Hugging Face's logo

hajeong67
/
llama-3.2-1B-instruct-lora

PEFT
TensorBoard
Safetensors
Model card Files Metrics Community
hajeong67 commited on
Commit
08f7e22
·
verified ·
1 Parent(s): eb89132

Upload folder using huggingface_hub

Browse files
Files changed (28) hide show
  1. adapter_config.json +2 -2
  2. adapter_model.safetensors +1 -1
  3. checkpoint-1400/README.md +202 -0
  4. checkpoint-1400/adapter_config.json +29 -0
  5. checkpoint-1400/adapter_model.safetensors +3 -0
  6. checkpoint-1400/optimizer.pt +3 -0
  7. checkpoint-1400/rng_state.pth +3 -0
  8. checkpoint-1400/scheduler.pt +3 -0
  9. checkpoint-1400/trainer_state.json +1021 -0
  10. checkpoint-1400/training_args.bin +3 -0
  11. checkpoint-2800/README.md +202 -0
  12. checkpoint-2800/adapter_config.json +29 -0
  13. checkpoint-2800/adapter_model.safetensors +3 -0
  14. checkpoint-2800/optimizer.pt +3 -0
  15. checkpoint-2800/rng_state.pth +3 -0
  16. checkpoint-2800/scheduler.pt +3 -0
  17. checkpoint-2800/trainer_state.json +2009 -0
  18. checkpoint-2800/training_args.bin +3 -0
  19. checkpoint-4200/README.md +202 -0
  20. checkpoint-4200/adapter_config.json +29 -0
  21. checkpoint-4200/adapter_model.safetensors +3 -0
  22. checkpoint-4200/optimizer.pt +3 -0
  23. checkpoint-4200/rng_state.pth +3 -0
  24. checkpoint-4200/scheduler.pt +3 -0
  25. checkpoint-4200/trainer_state.json +2997 -0
  26. checkpoint-4200/training_args.bin +3 -0
  27. tokenizer.json +1 -1
  28. training_args.bin +1 -1
adapter_config.json CHANGED
@@ -20,8 +20,8 @@
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
23
- "v_proj",
24
- "q_proj"
25
  ],
26
  "task_type": "CAUSAL_LM",
27
  "use_dora": false,
 
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
23
+ "q_proj",
24
+ "v_proj"
25
  ],
26
  "task_type": "CAUSAL_LM",
27
  "use_dora": false,
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:d2cd35d0baac967d5cd52ce3fffd9ce200841b230f5ec3f3a91c2d8bca2247d5
3
  size 3416264
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee723bec1693ba5b9d142a337187f000298d74d4c3f5e60a3af750f37abdd575
3
  size 3416264
checkpoint-1400/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Llama-3.2-1B-Instruct
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.13.2
checkpoint-1400/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "meta-llama/Llama-3.2-1B-Instruct",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "q_proj",
24
+ "v_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoint-1400/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0f220f122473f4485476f89d3e87d3726db5d7e92e67b9efc224f48788400e26
3
+ size 3416264
checkpoint-1400/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:88fde9cd38baef33ca74fb742c183f0779930bc9142b50da8f320a9cfe8f2107
3
+ size 6869818
checkpoint-1400/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6e654950e8ee96fc7cda636036720d303346a436a535ba00c826eceac02a302f
3
+ size 14244
checkpoint-1400/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:478347b0b3b6c101c738cc0036ec60acd3b507c9bbf381c9cd9dab2658156e04
3
+ size 1064
checkpoint-1400/trainer_state.json ADDED
@@ -0,0 +1,1021 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 1.0,
5
+ "eval_steps": 500,
6
+ "global_step": 1400,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.007142857142857143,
13
+ "grad_norm": 2.519733190536499,
14
+ "learning_rate": 9.976190476190477e-05,
15
+ "loss": 1.2999,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.014285714285714285,
20
+ "grad_norm": 2.4232161045074463,
21
+ "learning_rate": 9.952380952380953e-05,
22
+ "loss": 1.1228,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.02142857142857143,
27
+ "grad_norm": 2.7126998901367188,
28
+ "learning_rate": 9.92857142857143e-05,
29
+ "loss": 0.9913,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.02857142857142857,
34
+ "grad_norm": 7.3811235427856445,
35
+ "learning_rate": 9.904761904761905e-05,
36
+ "loss": 0.9306,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.03571428571428571,
41
+ "grad_norm": 6.5697197914123535,
42
+ "learning_rate": 9.880952380952381e-05,
43
+ "loss": 0.7457,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.04285714285714286,
48
+ "grad_norm": 1.2972687482833862,
49
+ "learning_rate": 9.857142857142858e-05,
50
+ "loss": 0.6408,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.05,
55
+ "grad_norm": 3.600625991821289,
56
+ "learning_rate": 9.833333333333333e-05,
57
+ "loss": 0.6541,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.05714285714285714,
62
+ "grad_norm": 1.5594621896743774,
63
+ "learning_rate": 9.80952380952381e-05,
64
+ "loss": 0.538,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.06428571428571428,
69
+ "grad_norm": 4.595429420471191,
70
+ "learning_rate": 9.785714285714286e-05,
71
+ "loss": 0.556,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.07142857142857142,
76
+ "grad_norm": 2.620643138885498,
77
+ "learning_rate": 9.761904761904762e-05,
78
+ "loss": 0.5521,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.07857142857142857,
83
+ "grad_norm": 3.503641366958618,
84
+ "learning_rate": 9.738095238095239e-05,
85
+ "loss": 0.5912,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.08571428571428572,
90
+ "grad_norm": 1.5058848857879639,
91
+ "learning_rate": 9.714285714285715e-05,
92
+ "loss": 0.5421,
93
+ "step": 120
94
+ },
95
+ {
96
+ "epoch": 0.09285714285714286,
97
+ "grad_norm": 3.7034990787506104,
98
+ "learning_rate": 9.69047619047619e-05,
99
+ "loss": 0.5568,
100
+ "step": 130
101
+ },
102
+ {
103
+ "epoch": 0.1,
104
+ "grad_norm": 3.714170455932617,
105
+ "learning_rate": 9.666666666666667e-05,
106
+ "loss": 0.5286,
107
+ "step": 140
108
+ },
109
+ {
110
+ "epoch": 0.10714285714285714,
111
+ "grad_norm": 2.069328546524048,
112
+ "learning_rate": 9.642857142857143e-05,
113
+ "loss": 0.4755,
114
+ "step": 150
115
+ },
116
+ {
117
+ "epoch": 0.11428571428571428,
118
+ "grad_norm": 1.2878583669662476,
119
+ "learning_rate": 9.61904761904762e-05,
120
+ "loss": 0.4817,
121
+ "step": 160
122
+ },
123
+ {
124
+ "epoch": 0.12142857142857143,
125
+ "grad_norm": 3.1945483684539795,
126
+ "learning_rate": 9.595238095238096e-05,
127
+ "loss": 0.5035,
128
+ "step": 170
129
+ },
130
+ {
131
+ "epoch": 0.12857142857142856,
132
+ "grad_norm": 1.9643166065216064,
133
+ "learning_rate": 9.571428571428573e-05,
134
+ "loss": 0.4932,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 0.1357142857142857,
139
+ "grad_norm": 3.5002968311309814,
140
+ "learning_rate": 9.547619047619049e-05,
141
+ "loss": 0.5531,
142
+ "step": 190
143
+ },
144
+ {
145
+ "epoch": 0.14285714285714285,
146
+ "grad_norm": 4.7095794677734375,
147
+ "learning_rate": 9.523809523809524e-05,
148
+ "loss": 0.4601,
149
+ "step": 200
150
+ },
151
+ {
152
+ "epoch": 0.15,
153
+ "grad_norm": 1.4850050210952759,
154
+ "learning_rate": 9.5e-05,
155
+ "loss": 0.432,
156
+ "step": 210
157
+ },
158
+ {
159
+ "epoch": 0.15714285714285714,
160
+ "grad_norm": 2.860565423965454,
161
+ "learning_rate": 9.476190476190476e-05,
162
+ "loss": 0.4699,
163
+ "step": 220
164
+ },
165
+ {
166
+ "epoch": 0.16428571428571428,
167
+ "grad_norm": 1.8467603921890259,
168
+ "learning_rate": 9.452380952380952e-05,
169
+ "loss": 0.42,
170
+ "step": 230
171
+ },
172
+ {
173
+ "epoch": 0.17142857142857143,
174
+ "grad_norm": 1.6448352336883545,
175
+ "learning_rate": 9.428571428571429e-05,
176
+ "loss": 0.431,
177
+ "step": 240
178
+ },
179
+ {
180
+ "epoch": 0.17857142857142858,
181
+ "grad_norm": 1.8279732465744019,
182
+ "learning_rate": 9.404761904761905e-05,
183
+ "loss": 0.4413,
184
+ "step": 250
185
+ },
186
+ {
187
+ "epoch": 0.18571428571428572,
188
+ "grad_norm": 3.0112857818603516,
189
+ "learning_rate": 9.380952380952381e-05,
190
+ "loss": 0.4263,
191
+ "step": 260
192
+ },
193
+ {
194
+ "epoch": 0.19285714285714287,
195
+ "grad_norm": 1.7717293500900269,
196
+ "learning_rate": 9.357142857142858e-05,
197
+ "loss": 0.4829,
198
+ "step": 270
199
+ },
200
+ {
201
+ "epoch": 0.2,
202
+ "grad_norm": 4.515851974487305,
203
+ "learning_rate": 9.333333333333334e-05,
204
+ "loss": 0.4194,
205
+ "step": 280
206
+ },
207
+ {
208
+ "epoch": 0.20714285714285716,
209
+ "grad_norm": 3.998619556427002,
210
+ "learning_rate": 9.309523809523811e-05,
211
+ "loss": 0.4422,
212
+ "step": 290
213
+ },
214
+ {
215
+ "epoch": 0.21428571428571427,
216
+ "grad_norm": 3.5681912899017334,
217
+ "learning_rate": 9.285714285714286e-05,
218
+ "loss": 0.4677,
219
+ "step": 300
220
+ },
221
+ {
222
+ "epoch": 0.22142857142857142,
223
+ "grad_norm": 3.286348342895508,
224
+ "learning_rate": 9.261904761904762e-05,
225
+ "loss": 0.4152,
226
+ "step": 310
227
+ },
228
+ {
229
+ "epoch": 0.22857142857142856,
230
+ "grad_norm": 2.018721580505371,
231
+ "learning_rate": 9.238095238095239e-05,
232
+ "loss": 0.5065,
233
+ "step": 320
234
+ },
235
+ {
236
+ "epoch": 0.2357142857142857,
237
+ "grad_norm": 1.8489034175872803,
238
+ "learning_rate": 9.214285714285714e-05,
239
+ "loss": 0.392,
240
+ "step": 330
241
+ },
242
+ {
243
+ "epoch": 0.24285714285714285,
244
+ "grad_norm": 1.3958369493484497,
245
+ "learning_rate": 9.19047619047619e-05,
246
+ "loss": 0.3988,
247
+ "step": 340
248
+ },
249
+ {
250
+ "epoch": 0.25,
251
+ "grad_norm": 2.9772961139678955,
252
+ "learning_rate": 9.166666666666667e-05,
253
+ "loss": 0.3955,
254
+ "step": 350
255
+ },
256
+ {
257
+ "epoch": 0.2571428571428571,
258
+ "grad_norm": 3.2752914428710938,
259
+ "learning_rate": 9.142857142857143e-05,
260
+ "loss": 0.3998,
261
+ "step": 360
262
+ },
263
+ {
264
+ "epoch": 0.2642857142857143,
265
+ "grad_norm": 1.8811057806015015,
266
+ "learning_rate": 9.11904761904762e-05,
267
+ "loss": 0.4775,
268
+ "step": 370
269
+ },
270
+ {
271
+ "epoch": 0.2714285714285714,
272
+ "grad_norm": 2.3310513496398926,
273
+ "learning_rate": 9.095238095238096e-05,
274
+ "loss": 0.3703,
275
+ "step": 380
276
+ },
277
+ {
278
+ "epoch": 0.2785714285714286,
279
+ "grad_norm": 2.967597484588623,
280
+ "learning_rate": 9.071428571428571e-05,
281
+ "loss": 0.4177,
282
+ "step": 390
283
+ },
284
+ {
285
+ "epoch": 0.2857142857142857,
286
+ "grad_norm": 2.671673536300659,
287
+ "learning_rate": 9.047619047619048e-05,
288
+ "loss": 0.3823,
289
+ "step": 400
290
+ },
291
+ {
292
+ "epoch": 0.29285714285714287,
293
+ "grad_norm": 4.647867679595947,
294
+ "learning_rate": 9.023809523809524e-05,
295
+ "loss": 0.3888,
296
+ "step": 410
297
+ },
298
+ {
299
+ "epoch": 0.3,
300
+ "grad_norm": 3.4788432121276855,
301
+ "learning_rate": 9e-05,
302
+ "loss": 0.3856,
303
+ "step": 420
304
+ },
305
+ {
306
+ "epoch": 0.30714285714285716,
307
+ "grad_norm": 3.308420181274414,
308
+ "learning_rate": 8.976190476190477e-05,
309
+ "loss": 0.4143,
310
+ "step": 430
311
+ },
312
+ {
313
+ "epoch": 0.3142857142857143,
314
+ "grad_norm": 3.499880075454712,
315
+ "learning_rate": 8.952380952380953e-05,
316
+ "loss": 0.4857,
317
+ "step": 440
318
+ },
319
+ {
320
+ "epoch": 0.32142857142857145,
321
+ "grad_norm": 1.7907614707946777,
322
+ "learning_rate": 8.92857142857143e-05,
323
+ "loss": 0.356,
324
+ "step": 450
325
+ },
326
+ {
327
+ "epoch": 0.32857142857142857,
328
+ "grad_norm": 3.0111501216888428,
329
+ "learning_rate": 8.904761904761905e-05,
330
+ "loss": 0.358,
331
+ "step": 460
332
+ },
333
+ {
334
+ "epoch": 0.3357142857142857,
335
+ "grad_norm": 2.1389689445495605,
336
+ "learning_rate": 8.880952380952381e-05,
337
+ "loss": 0.3331,
338
+ "step": 470
339
+ },
340
+ {
341
+ "epoch": 0.34285714285714286,
342
+ "grad_norm": 1.8007326126098633,
343
+ "learning_rate": 8.857142857142857e-05,
344
+ "loss": 0.3148,
345
+ "step": 480
346
+ },
347
+ {
348
+ "epoch": 0.35,
349
+ "grad_norm": 3.603548288345337,
350
+ "learning_rate": 8.833333333333333e-05,
351
+ "loss": 0.4513,
352
+ "step": 490
353
+ },
354
+ {
355
+ "epoch": 0.35714285714285715,
356
+ "grad_norm": 2.270984411239624,
357
+ "learning_rate": 8.80952380952381e-05,
358
+ "loss": 0.3392,
359
+ "step": 500
360
+ },
361
+ {
362
+ "epoch": 0.36428571428571427,
363
+ "grad_norm": 1.5367858409881592,
364
+ "learning_rate": 8.785714285714286e-05,
365
+ "loss": 0.3665,
366
+ "step": 510
367
+ },
368
+ {
369
+ "epoch": 0.37142857142857144,
370
+ "grad_norm": 1.5397262573242188,
371
+ "learning_rate": 8.761904761904762e-05,
372
+ "loss": 0.3875,
373
+ "step": 520
374
+ },
375
+ {
376
+ "epoch": 0.37857142857142856,
377
+ "grad_norm": 3.6661577224731445,
378
+ "learning_rate": 8.738095238095239e-05,
379
+ "loss": 0.3617,
380
+ "step": 530
381
+ },
382
+ {
383
+ "epoch": 0.38571428571428573,
384
+ "grad_norm": 2.5946245193481445,
385
+ "learning_rate": 8.714285714285715e-05,
386
+ "loss": 0.3318,
387
+ "step": 540
388
+ },
389
+ {
390
+ "epoch": 0.39285714285714285,
391
+ "grad_norm": 2.508056640625,
392
+ "learning_rate": 8.690476190476192e-05,
393
+ "loss": 0.3681,
394
+ "step": 550
395
+ },
396
+ {
397
+ "epoch": 0.4,
398
+ "grad_norm": 5.406072616577148,
399
+ "learning_rate": 8.666666666666667e-05,
400
+ "loss": 0.3589,
401
+ "step": 560
402
+ },
403
+ {
404
+ "epoch": 0.40714285714285714,
405
+ "grad_norm": 2.0363543033599854,
406
+ "learning_rate": 8.642857142857143e-05,
407
+ "loss": 0.4328,
408
+ "step": 570
409
+ },
410
+ {
411
+ "epoch": 0.4142857142857143,
412
+ "grad_norm": 1.796236515045166,
413
+ "learning_rate": 8.61904761904762e-05,
414
+ "loss": 0.4005,
415
+ "step": 580
416
+ },
417
+ {
418
+ "epoch": 0.42142857142857143,
419
+ "grad_norm": 5.6172943115234375,
420
+ "learning_rate": 8.595238095238096e-05,
421
+ "loss": 0.3414,
422
+ "step": 590
423
+ },
424
+ {
425
+ "epoch": 0.42857142857142855,
426
+ "grad_norm": 2.953678846359253,
427
+ "learning_rate": 8.571428571428571e-05,
428
+ "loss": 0.3917,
429
+ "step": 600
430
+ },
431
+ {
432
+ "epoch": 0.4357142857142857,
433
+ "grad_norm": 1.8276687860488892,
434
+ "learning_rate": 8.547619047619048e-05,
435
+ "loss": 0.4057,
436
+ "step": 610
437
+ },
438
+ {
439
+ "epoch": 0.44285714285714284,
440
+ "grad_norm": 1.8668811321258545,
441
+ "learning_rate": 8.523809523809524e-05,
442
+ "loss": 0.4153,
443
+ "step": 620
444
+ },
445
+ {
446
+ "epoch": 0.45,
447
+ "grad_norm": 2.87809419631958,
448
+ "learning_rate": 8.5e-05,
449
+ "loss": 0.3911,
450
+ "step": 630
451
+ },
452
+ {
453
+ "epoch": 0.45714285714285713,
454
+ "grad_norm": 4.863674640655518,
455
+ "learning_rate": 8.476190476190477e-05,
456
+ "loss": 0.4458,
457
+ "step": 640
458
+ },
459
+ {
460
+ "epoch": 0.4642857142857143,
461
+ "grad_norm": 1.7150845527648926,
462
+ "learning_rate": 8.452380952380952e-05,
463
+ "loss": 0.3951,
464
+ "step": 650
465
+ },
466
+ {
467
+ "epoch": 0.4714285714285714,
468
+ "grad_norm": 2.0149145126342773,
469
+ "learning_rate": 8.428571428571429e-05,
470
+ "loss": 0.3238,
471
+ "step": 660
472
+ },
473
+ {
474
+ "epoch": 0.4785714285714286,
475
+ "grad_norm": 2.3146402835845947,
476
+ "learning_rate": 8.404761904761905e-05,
477
+ "loss": 0.3129,
478
+ "step": 670
479
+ },
480
+ {
481
+ "epoch": 0.4857142857142857,
482
+ "grad_norm": 1.8723537921905518,
483
+ "learning_rate": 8.380952380952382e-05,
484
+ "loss": 0.2969,
485
+ "step": 680
486
+ },
487
+ {
488
+ "epoch": 0.4928571428571429,
489
+ "grad_norm": 1.6206679344177246,
490
+ "learning_rate": 8.357142857142858e-05,
491
+ "loss": 0.3673,
492
+ "step": 690
493
+ },
494
+ {
495
+ "epoch": 0.5,
496
+ "grad_norm": 2.490722179412842,
497
+ "learning_rate": 8.333333333333334e-05,
498
+ "loss": 0.3493,
499
+ "step": 700
500
+ },
501
+ {
502
+ "epoch": 0.5071428571428571,
503
+ "grad_norm": 1.7868791818618774,
504
+ "learning_rate": 8.309523809523811e-05,
505
+ "loss": 0.423,
506
+ "step": 710
507
+ },
508
+ {
509
+ "epoch": 0.5142857142857142,
510
+ "grad_norm": 2.3152015209198,
511
+ "learning_rate": 8.285714285714287e-05,
512
+ "loss": 0.3448,
513
+ "step": 720
514
+ },
515
+ {
516
+ "epoch": 0.5214285714285715,
517
+ "grad_norm": 1.934523105621338,
518
+ "learning_rate": 8.261904761904762e-05,
519
+ "loss": 0.372,
520
+ "step": 730
521
+ },
522
+ {
523
+ "epoch": 0.5285714285714286,
524
+ "grad_norm": 2.421624183654785,
525
+ "learning_rate": 8.238095238095238e-05,
526
+ "loss": 0.3602,
527
+ "step": 740
528
+ },
529
+ {
530
+ "epoch": 0.5357142857142857,
531
+ "grad_norm": 2.0156285762786865,
532
+ "learning_rate": 8.214285714285714e-05,
533
+ "loss": 0.334,
534
+ "step": 750
535
+ },
536
+ {
537
+ "epoch": 0.5428571428571428,
538
+ "grad_norm": 1.6513855457305908,
539
+ "learning_rate": 8.19047619047619e-05,
540
+ "loss": 0.3553,
541
+ "step": 760
542
+ },
543
+ {
544
+ "epoch": 0.55,
545
+ "grad_norm": 1.950451374053955,
546
+ "learning_rate": 8.166666666666667e-05,
547
+ "loss": 0.3696,
548
+ "step": 770
549
+ },
550
+ {
551
+ "epoch": 0.5571428571428572,
552
+ "grad_norm": 2.4662423133850098,
553
+ "learning_rate": 8.142857142857143e-05,
554
+ "loss": 0.3232,
555
+ "step": 780
556
+ },
557
+ {
558
+ "epoch": 0.5642857142857143,
559
+ "grad_norm": 2.060654401779175,
560
+ "learning_rate": 8.11904761904762e-05,
561
+ "loss": 0.3294,
562
+ "step": 790
563
+ },
564
+ {
565
+ "epoch": 0.5714285714285714,
566
+ "grad_norm": 3.3581550121307373,
567
+ "learning_rate": 8.095238095238096e-05,
568
+ "loss": 0.3022,
569
+ "step": 800
570
+ },
571
+ {
572
+ "epoch": 0.5785714285714286,
573
+ "grad_norm": 1.6778148412704468,
574
+ "learning_rate": 8.071428571428573e-05,
575
+ "loss": 0.2808,
576
+ "step": 810
577
+ },
578
+ {
579
+ "epoch": 0.5857142857142857,
580
+ "grad_norm": 5.509958267211914,
581
+ "learning_rate": 8.047619047619048e-05,
582
+ "loss": 0.3584,
583
+ "step": 820
584
+ },
585
+ {
586
+ "epoch": 0.5928571428571429,
587
+ "grad_norm": 2.3136510848999023,
588
+ "learning_rate": 8.023809523809524e-05,
589
+ "loss": 0.2967,
590
+ "step": 830
591
+ },
592
+ {
593
+ "epoch": 0.6,
594
+ "grad_norm": 2.757749557495117,
595
+ "learning_rate": 8e-05,
596
+ "loss": 0.3091,
597
+ "step": 840
598
+ },
599
+ {
600
+ "epoch": 0.6071428571428571,
601
+ "grad_norm": 1.6453288793563843,
602
+ "learning_rate": 7.976190476190477e-05,
603
+ "loss": 0.2882,
604
+ "step": 850
605
+ },
606
+ {
607
+ "epoch": 0.6142857142857143,
608
+ "grad_norm": 3.599712371826172,
609
+ "learning_rate": 7.952380952380952e-05,
610
+ "loss": 0.35,
611
+ "step": 860
612
+ },
613
+ {
614
+ "epoch": 0.6214285714285714,
615
+ "grad_norm": 1.203050136566162,
616
+ "learning_rate": 7.928571428571429e-05,
617
+ "loss": 0.3147,
618
+ "step": 870
619
+ },
620
+ {
621
+ "epoch": 0.6285714285714286,
622
+ "grad_norm": 3.0582826137542725,
623
+ "learning_rate": 7.904761904761905e-05,
624
+ "loss": 0.369,
625
+ "step": 880
626
+ },
627
+ {
628
+ "epoch": 0.6357142857142857,
629
+ "grad_norm": 1.7133578062057495,
630
+ "learning_rate": 7.880952380952382e-05,
631
+ "loss": 0.4404,
632
+ "step": 890
633
+ },
634
+ {
635
+ "epoch": 0.6428571428571429,
636
+ "grad_norm": 2.1532819271087646,
637
+ "learning_rate": 7.857142857142858e-05,
638
+ "loss": 0.3249,
639
+ "step": 900
640
+ },
641
+ {
642
+ "epoch": 0.65,
643
+ "grad_norm": 3.529282808303833,
644
+ "learning_rate": 7.833333333333333e-05,
645
+ "loss": 0.3449,
646
+ "step": 910
647
+ },
648
+ {
649
+ "epoch": 0.6571428571428571,
650
+ "grad_norm": 3.412729501724243,
651
+ "learning_rate": 7.80952380952381e-05,
652
+ "loss": 0.3678,
653
+ "step": 920
654
+ },
655
+ {
656
+ "epoch": 0.6642857142857143,
657
+ "grad_norm": 1.8676691055297852,
658
+ "learning_rate": 7.785714285714286e-05,
659
+ "loss": 0.3522,
660
+ "step": 930
661
+ },
662
+ {
663
+ "epoch": 0.6714285714285714,
664
+ "grad_norm": 1.8748834133148193,
665
+ "learning_rate": 7.761904761904762e-05,
666
+ "loss": 0.2561,
667
+ "step": 940
668
+ },
669
+ {
670
+ "epoch": 0.6785714285714286,
671
+ "grad_norm": 3.1622543334960938,
672
+ "learning_rate": 7.738095238095239e-05,
673
+ "loss": 0.3245,
674
+ "step": 950
675
+ },
676
+ {
677
+ "epoch": 0.6857142857142857,
678
+ "grad_norm": 2.17108416557312,
679
+ "learning_rate": 7.714285714285715e-05,
680
+ "loss": 0.335,
681
+ "step": 960
682
+ },
683
+ {
684
+ "epoch": 0.6928571428571428,
685
+ "grad_norm": 3.686673402786255,
686
+ "learning_rate": 7.690476190476192e-05,
687
+ "loss": 0.319,
688
+ "step": 970
689
+ },
690
+ {
691
+ "epoch": 0.7,
692
+ "grad_norm": 7.108774662017822,
693
+ "learning_rate": 7.666666666666667e-05,
694
+ "loss": 0.3337,
695
+ "step": 980
696
+ },
697
+ {
698
+ "epoch": 0.7071428571428572,
699
+ "grad_norm": 2.532336711883545,
700
+ "learning_rate": 7.642857142857143e-05,
701
+ "loss": 0.377,
702
+ "step": 990
703
+ },
704
+ {
705
+ "epoch": 0.7142857142857143,
706
+ "grad_norm": 1.8768038749694824,
707
+ "learning_rate": 7.619047619047618e-05,
708
+ "loss": 0.3745,
709
+ "step": 1000
710
+ },
711
+ {
712
+ "epoch": 0.7214285714285714,
713
+ "grad_norm": 1.8926101922988892,
714
+ "learning_rate": 7.595238095238095e-05,
715
+ "loss": 0.3503,
716
+ "step": 1010
717
+ },
718
+ {
719
+ "epoch": 0.7285714285714285,
720
+ "grad_norm": 2.2763147354125977,
721
+ "learning_rate": 7.571428571428571e-05,
722
+ "loss": 0.307,
723
+ "step": 1020
724
+ },
725
+ {
726
+ "epoch": 0.7357142857142858,
727
+ "grad_norm": 3.663257360458374,
728
+ "learning_rate": 7.547619047619048e-05,
729
+ "loss": 0.2876,
730
+ "step": 1030
731
+ },
732
+ {
733
+ "epoch": 0.7428571428571429,
734
+ "grad_norm": 1.844648003578186,
735
+ "learning_rate": 7.523809523809524e-05,
736
+ "loss": 0.3512,
737
+ "step": 1040
738
+ },
739
+ {
740
+ "epoch": 0.75,
741
+ "grad_norm": 2.5800540447235107,
742
+ "learning_rate": 7.500000000000001e-05,
743
+ "loss": 0.301,
744
+ "step": 1050
745
+ },
746
+ {
747
+ "epoch": 0.7571428571428571,
748
+ "grad_norm": 2.4674232006073,
749
+ "learning_rate": 7.476190476190477e-05,
750
+ "loss": 0.3635,
751
+ "step": 1060
752
+ },
753
+ {
754
+ "epoch": 0.7642857142857142,
755
+ "grad_norm": 3.5119080543518066,
756
+ "learning_rate": 7.452380952380952e-05,
757
+ "loss": 0.3688,
758
+ "step": 1070
759
+ },
760
+ {
761
+ "epoch": 0.7714285714285715,
762
+ "grad_norm": 2.5902345180511475,
763
+ "learning_rate": 7.428571428571429e-05,
764
+ "loss": 0.3136,
765
+ "step": 1080
766
+ },
767
+ {
768
+ "epoch": 0.7785714285714286,
769
+ "grad_norm": 1.734082818031311,
770
+ "learning_rate": 7.404761904761905e-05,
771
+ "loss": 0.3366,
772
+ "step": 1090
773
+ },
774
+ {
775
+ "epoch": 0.7857142857142857,
776
+ "grad_norm": 2.0863406658172607,
777
+ "learning_rate": 7.380952380952382e-05,
778
+ "loss": 0.331,
779
+ "step": 1100
780
+ },
781
+ {
782
+ "epoch": 0.7928571428571428,
783
+ "grad_norm": 2.9227652549743652,
784
+ "learning_rate": 7.357142857142858e-05,
785
+ "loss": 0.3131,
786
+ "step": 1110
787
+ },
788
+ {
789
+ "epoch": 0.8,
790
+ "grad_norm": 5.268605709075928,
791
+ "learning_rate": 7.333333333333333e-05,
792
+ "loss": 0.3212,
793
+ "step": 1120
794
+ },
795
+ {
796
+ "epoch": 0.8071428571428572,
797
+ "grad_norm": 2.633431911468506,
798
+ "learning_rate": 7.30952380952381e-05,
799
+ "loss": 0.2707,
800
+ "step": 1130
801
+ },
802
+ {
803
+ "epoch": 0.8142857142857143,
804
+ "grad_norm": 1.652354121208191,
805
+ "learning_rate": 7.285714285714286e-05,
806
+ "loss": 0.3037,
807
+ "step": 1140
808
+ },
809
+ {
810
+ "epoch": 0.8214285714285714,
811
+ "grad_norm": 3.1987786293029785,
812
+ "learning_rate": 7.261904761904762e-05,
813
+ "loss": 0.3603,
814
+ "step": 1150
815
+ },
816
+ {
817
+ "epoch": 0.8285714285714286,
818
+ "grad_norm": 1.7624555826187134,
819
+ "learning_rate": 7.238095238095238e-05,
820
+ "loss": 0.308,
821
+ "step": 1160
822
+ },
823
+ {
824
+ "epoch": 0.8357142857142857,
825
+ "grad_norm": 2.6379318237304688,
826
+ "learning_rate": 7.214285714285714e-05,
827
+ "loss": 0.3068,
828
+ "step": 1170
829
+ },
830
+ {
831
+ "epoch": 0.8428571428571429,
832
+ "grad_norm": 2.6688101291656494,
833
+ "learning_rate": 7.19047619047619e-05,
834
+ "loss": 0.3134,
835
+ "step": 1180
836
+ },
837
+ {
838
+ "epoch": 0.85,
839
+ "grad_norm": 3.698643207550049,
840
+ "learning_rate": 7.166666666666667e-05,
841
+ "loss": 0.2973,
842
+ "step": 1190
843
+ },
844
+ {
845
+ "epoch": 0.8571428571428571,
846
+ "grad_norm": 2.9004178047180176,
847
+ "learning_rate": 7.142857142857143e-05,
848
+ "loss": 0.3116,
849
+ "step": 1200
850
+ },
851
+ {
852
+ "epoch": 0.8642857142857143,
853
+ "grad_norm": 2.3414883613586426,
854
+ "learning_rate": 7.11904761904762e-05,
855
+ "loss": 0.334,
856
+ "step": 1210
857
+ },
858
+ {
859
+ "epoch": 0.8714285714285714,
860
+ "grad_norm": 3.6203651428222656,
861
+ "learning_rate": 7.095238095238096e-05,
862
+ "loss": 0.3241,
863
+ "step": 1220
864
+ },
865
+ {
866
+ "epoch": 0.8785714285714286,
867
+ "grad_norm": 2.7472829818725586,
868
+ "learning_rate": 7.071428571428573e-05,
869
+ "loss": 0.2727,
870
+ "step": 1230
871
+ },
872
+ {
873
+ "epoch": 0.8857142857142857,
874
+ "grad_norm": 3.2332115173339844,
875
+ "learning_rate": 7.047619047619048e-05,
876
+ "loss": 0.3329,
877
+ "step": 1240
878
+ },
879
+ {
880
+ "epoch": 0.8928571428571429,
881
+ "grad_norm": 1.968595027923584,
882
+ "learning_rate": 7.023809523809524e-05,
883
+ "loss": 0.3478,
884
+ "step": 1250
885
+ },
886
+ {
887
+ "epoch": 0.9,
888
+ "grad_norm": 2.3712966442108154,
889
+ "learning_rate": 7e-05,
890
+ "loss": 0.3964,
891
+ "step": 1260
892
+ },
893
+ {
894
+ "epoch": 0.9071428571428571,
895
+ "grad_norm": 1.7704048156738281,
896
+ "learning_rate": 6.976190476190476e-05,
897
+ "loss": 0.3983,
898
+ "step": 1270
899
+ },
900
+ {
901
+ "epoch": 0.9142857142857143,
902
+ "grad_norm": 1.8879201412200928,
903
+ "learning_rate": 6.952380952380952e-05,
904
+ "loss": 0.3296,
905
+ "step": 1280
906
+ },
907
+ {
908
+ "epoch": 0.9214285714285714,
909
+ "grad_norm": 4.0522780418396,
910
+ "learning_rate": 6.928571428571429e-05,
911
+ "loss": 0.3722,
912
+ "step": 1290
913
+ },
914
+ {
915
+ "epoch": 0.9285714285714286,
916
+ "grad_norm": 1.7849159240722656,
917
+ "learning_rate": 6.904761904761905e-05,
918
+ "loss": 0.3022,
919
+ "step": 1300
920
+ },
921
+ {
922
+ "epoch": 0.9357142857142857,
923
+ "grad_norm": 3.376678705215454,
924
+ "learning_rate": 6.880952380952382e-05,
925
+ "loss": 0.3429,
926
+ "step": 1310
927
+ },
928
+ {
929
+ "epoch": 0.9428571428571428,
930
+ "grad_norm": 4.100617408752441,
931
+ "learning_rate": 6.857142857142858e-05,
932
+ "loss": 0.2793,
933
+ "step": 1320
934
+ },
935
+ {
936
+ "epoch": 0.95,
937
+ "grad_norm": 3.8000288009643555,
938
+ "learning_rate": 6.833333333333333e-05,
939
+ "loss": 0.3114,
940
+ "step": 1330
941
+ },
942
+ {
943
+ "epoch": 0.9571428571428572,
944
+ "grad_norm": 4.846517562866211,
945
+ "learning_rate": 6.80952380952381e-05,
946
+ "loss": 0.3377,
947
+ "step": 1340
948
+ },
949
+ {
950
+ "epoch": 0.9642857142857143,
951
+ "grad_norm": 2.0567409992218018,
952
+ "learning_rate": 6.785714285714286e-05,
953
+ "loss": 0.3198,
954
+ "step": 1350
955
+ },
956
+ {
957
+ "epoch": 0.9714285714285714,
958
+ "grad_norm": 2.3197948932647705,
959
+ "learning_rate": 6.761904761904763e-05,
960
+ "loss": 0.2817,
961
+ "step": 1360
962
+ },
963
+ {
964
+ "epoch": 0.9785714285714285,
965
+ "grad_norm": 2.7312827110290527,
966
+ "learning_rate": 6.738095238095239e-05,
967
+ "loss": 0.3378,
968
+ "step": 1370
969
+ },
970
+ {
971
+ "epoch": 0.9857142857142858,
972
+ "grad_norm": 3.115516185760498,
973
+ "learning_rate": 6.714285714285714e-05,
974
+ "loss": 0.3638,
975
+ "step": 1380
976
+ },
977
+ {
978
+ "epoch": 0.9928571428571429,
979
+ "grad_norm": 1.0742294788360596,
980
+ "learning_rate": 6.69047619047619e-05,
981
+ "loss": 0.3176,
982
+ "step": 1390
983
+ },
984
+ {
985
+ "epoch": 1.0,
986
+ "grad_norm": 2.3872761726379395,
987
+ "learning_rate": 6.666666666666667e-05,
988
+ "loss": 0.2921,
989
+ "step": 1400
990
+ },
991
+ {
992
+ "epoch": 1.0,
993
+ "eval_loss": 0.2929375469684601,
994
+ "eval_runtime": 68.5646,
995
+ "eval_samples_per_second": 2.917,
996
+ "eval_steps_per_second": 0.365,
997
+ "step": 1400
998
+ }
999
+ ],
1000
+ "logging_steps": 10,
1001
+ "max_steps": 4200,
1002
+ "num_input_tokens_seen": 0,
1003
+ "num_train_epochs": 3,
1004
+ "save_steps": 500,
1005
+ "stateful_callbacks": {
1006
+ "TrainerControl": {
1007
+ "args": {
1008
+ "should_epoch_stop": false,
1009
+ "should_evaluate": false,
1010
+ "should_log": false,
1011
+ "should_save": true,
1012
+ "should_training_stop": false
1013
+ },
1014
+ "attributes": {}
1015
+ }
1016
+ },
1017
+ "total_flos": 8377941884928000.0,
1018
+ "train_batch_size": 2,
1019
+ "trial_name": null,
1020
+ "trial_params": null
1021
+ }
checkpoint-1400/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aef410e0e1d0110bb1523d95ae4f80f726beb1aea9164ada6672ae6ffeddc06b
3
+ size 5240
checkpoint-2800/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Llama-3.2-1B-Instruct
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.13.2
checkpoint-2800/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "meta-llama/Llama-3.2-1B-Instruct",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "q_proj",
24
+ "v_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoint-2800/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:93da248a9807da542579e072ee689ee2a50eb155c2e74d2c2ed70e024ed2f182
3
+ size 3416264
checkpoint-2800/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4a40388f2d838cdd1692203bac646cc7d182cfbb33ac089c3bcd395d42a2f25a
3
+ size 6869818
checkpoint-2800/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:36f3f5b1fa8e2680054bd92408a26f568f9b1142bccee53ebe01cc7a97298abe
3
+ size 14244
checkpoint-2800/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f33a5a58f0d3724566602ae29967ea8fec117026be735d4fddb564a5e2204675
3
+ size 1064
checkpoint-2800/trainer_state.json ADDED
@@ -0,0 +1,2009 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 2.0,
5
+ "eval_steps": 500,
6
+ "global_step": 2800,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.007142857142857143,
13
+ "grad_norm": 2.519733190536499,
14
+ "learning_rate": 9.976190476190477e-05,
15
+ "loss": 1.2999,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.014285714285714285,
20
+ "grad_norm": 2.4232161045074463,
21
+ "learning_rate": 9.952380952380953e-05,
22
+ "loss": 1.1228,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.02142857142857143,
27
+ "grad_norm": 2.7126998901367188,
28
+ "learning_rate": 9.92857142857143e-05,
29
+ "loss": 0.9913,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.02857142857142857,
34
+ "grad_norm": 7.3811235427856445,
35
+ "learning_rate": 9.904761904761905e-05,
36
+ "loss": 0.9306,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.03571428571428571,
41
+ "grad_norm": 6.5697197914123535,
42
+ "learning_rate": 9.880952380952381e-05,
43
+ "loss": 0.7457,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.04285714285714286,
48
+ "grad_norm": 1.2972687482833862,
49
+ "learning_rate": 9.857142857142858e-05,
50
+ "loss": 0.6408,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.05,
55
+ "grad_norm": 3.600625991821289,
56
+ "learning_rate": 9.833333333333333e-05,
57
+ "loss": 0.6541,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.05714285714285714,
62
+ "grad_norm": 1.5594621896743774,
63
+ "learning_rate": 9.80952380952381e-05,
64
+ "loss": 0.538,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.06428571428571428,
69
+ "grad_norm": 4.595429420471191,
70
+ "learning_rate": 9.785714285714286e-05,
71
+ "loss": 0.556,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.07142857142857142,
76
+ "grad_norm": 2.620643138885498,
77
+ "learning_rate": 9.761904761904762e-05,
78
+ "loss": 0.5521,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.07857142857142857,
83
+ "grad_norm": 3.503641366958618,
84
+ "learning_rate": 9.738095238095239e-05,
85
+ "loss": 0.5912,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.08571428571428572,
90
+ "grad_norm": 1.5058848857879639,
91
+ "learning_rate": 9.714285714285715e-05,
92
+ "loss": 0.5421,
93
+ "step": 120
94
+ },
95
+ {
96
+ "epoch": 0.09285714285714286,
97
+ "grad_norm": 3.7034990787506104,
98
+ "learning_rate": 9.69047619047619e-05,
99
+ "loss": 0.5568,
100
+ "step": 130
101
+ },
102
+ {
103
+ "epoch": 0.1,
104
+ "grad_norm": 3.714170455932617,
105
+ "learning_rate": 9.666666666666667e-05,
106
+ "loss": 0.5286,
107
+ "step": 140
108
+ },
109
+ {
110
+ "epoch": 0.10714285714285714,
111
+ "grad_norm": 2.069328546524048,
112
+ "learning_rate": 9.642857142857143e-05,
113
+ "loss": 0.4755,
114
+ "step": 150
115
+ },
116
+ {
117
+ "epoch": 0.11428571428571428,
118
+ "grad_norm": 1.2878583669662476,
119
+ "learning_rate": 9.61904761904762e-05,
120
+ "loss": 0.4817,
121
+ "step": 160
122
+ },
123
+ {
124
+ "epoch": 0.12142857142857143,
125
+ "grad_norm": 3.1945483684539795,
126
+ "learning_rate": 9.595238095238096e-05,
127
+ "loss": 0.5035,
128
+ "step": 170
129
+ },
130
+ {
131
+ "epoch": 0.12857142857142856,
132
+ "grad_norm": 1.9643166065216064,
133
+ "learning_rate": 9.571428571428573e-05,
134
+ "loss": 0.4932,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 0.1357142857142857,
139
+ "grad_norm": 3.5002968311309814,
140
+ "learning_rate": 9.547619047619049e-05,
141
+ "loss": 0.5531,
142
+ "step": 190
143
+ },
144
+ {
145
+ "epoch": 0.14285714285714285,
146
+ "grad_norm": 4.7095794677734375,
147
+ "learning_rate": 9.523809523809524e-05,
148
+ "loss": 0.4601,
149
+ "step": 200
150
+ },
151
+ {
152
+ "epoch": 0.15,
153
+ "grad_norm": 1.4850050210952759,
154
+ "learning_rate": 9.5e-05,
155
+ "loss": 0.432,
156
+ "step": 210
157
+ },
158
+ {
159
+ "epoch": 0.15714285714285714,
160
+ "grad_norm": 2.860565423965454,
161
+ "learning_rate": 9.476190476190476e-05,
162
+ "loss": 0.4699,
163
+ "step": 220
164
+ },
165
+ {
166
+ "epoch": 0.16428571428571428,
167
+ "grad_norm": 1.8467603921890259,
168
+ "learning_rate": 9.452380952380952e-05,
169
+ "loss": 0.42,
170
+ "step": 230
171
+ },
172
+ {
173
+ "epoch": 0.17142857142857143,
174
+ "grad_norm": 1.6448352336883545,
175
+ "learning_rate": 9.428571428571429e-05,
176
+ "loss": 0.431,
177
+ "step": 240
178
+ },
179
+ {
180
+ "epoch": 0.17857142857142858,
181
+ "grad_norm": 1.8279732465744019,
182
+ "learning_rate": 9.404761904761905e-05,
183
+ "loss": 0.4413,
184
+ "step": 250
185
+ },
186
+ {
187
+ "epoch": 0.18571428571428572,
188
+ "grad_norm": 3.0112857818603516,
189
+ "learning_rate": 9.380952380952381e-05,
190
+ "loss": 0.4263,
191
+ "step": 260
192
+ },
193
+ {
194
+ "epoch": 0.19285714285714287,
195
+ "grad_norm": 1.7717293500900269,
196
+ "learning_rate": 9.357142857142858e-05,
197
+ "loss": 0.4829,
198
+ "step": 270
199
+ },
200
+ {
201
+ "epoch": 0.2,
202
+ "grad_norm": 4.515851974487305,
203
+ "learning_rate": 9.333333333333334e-05,
204
+ "loss": 0.4194,
205
+ "step": 280
206
+ },
207
+ {
208
+ "epoch": 0.20714285714285716,
209
+ "grad_norm": 3.998619556427002,
210
+ "learning_rate": 9.309523809523811e-05,
211
+ "loss": 0.4422,
212
+ "step": 290
213
+ },
214
+ {
215
+ "epoch": 0.21428571428571427,
216
+ "grad_norm": 3.5681912899017334,
217
+ "learning_rate": 9.285714285714286e-05,
218
+ "loss": 0.4677,
219
+ "step": 300
220
+ },
221
+ {
222
+ "epoch": 0.22142857142857142,
223
+ "grad_norm": 3.286348342895508,
224
+ "learning_rate": 9.261904761904762e-05,
225
+ "loss": 0.4152,
226
+ "step": 310
227
+ },
228
+ {
229
+ "epoch": 0.22857142857142856,
230
+ "grad_norm": 2.018721580505371,
231
+ "learning_rate": 9.238095238095239e-05,
232
+ "loss": 0.5065,
233
+ "step": 320
234
+ },
235
+ {
236
+ "epoch": 0.2357142857142857,
237
+ "grad_norm": 1.8489034175872803,
238
+ "learning_rate": 9.214285714285714e-05,
239
+ "loss": 0.392,
240
+ "step": 330
241
+ },
242
+ {
243
+ "epoch": 0.24285714285714285,
244
+ "grad_norm": 1.3958369493484497,
245
+ "learning_rate": 9.19047619047619e-05,
246
+ "loss": 0.3988,
247
+ "step": 340
248
+ },
249
+ {
250
+ "epoch": 0.25,
251
+ "grad_norm": 2.9772961139678955,
252
+ "learning_rate": 9.166666666666667e-05,
253
+ "loss": 0.3955,
254
+ "step": 350
255
+ },
256
+ {
257
+ "epoch": 0.2571428571428571,
258
+ "grad_norm": 3.2752914428710938,
259
+ "learning_rate": 9.142857142857143e-05,
260
+ "loss": 0.3998,
261
+ "step": 360
262
+ },
263
+ {
264
+ "epoch": 0.2642857142857143,
265
+ "grad_norm": 1.8811057806015015,
266
+ "learning_rate": 9.11904761904762e-05,
267
+ "loss": 0.4775,
268
+ "step": 370
269
+ },
270
+ {
271
+ "epoch": 0.2714285714285714,
272
+ "grad_norm": 2.3310513496398926,
273
+ "learning_rate": 9.095238095238096e-05,
274
+ "loss": 0.3703,
275
+ "step": 380
276
+ },
277
+ {
278
+ "epoch": 0.2785714285714286,
279
+ "grad_norm": 2.967597484588623,
280
+ "learning_rate": 9.071428571428571e-05,
281
+ "loss": 0.4177,
282
+ "step": 390
283
+ },
284
+ {
285
+ "epoch": 0.2857142857142857,
286
+ "grad_norm": 2.671673536300659,
287
+ "learning_rate": 9.047619047619048e-05,
288
+ "loss": 0.3823,
289
+ "step": 400
290
+ },
291
+ {
292
+ "epoch": 0.29285714285714287,
293
+ "grad_norm": 4.647867679595947,
294
+ "learning_rate": 9.023809523809524e-05,
295
+ "loss": 0.3888,
296
+ "step": 410
297
+ },
298
+ {
299
+ "epoch": 0.3,
300
+ "grad_norm": 3.4788432121276855,
301
+ "learning_rate": 9e-05,
302
+ "loss": 0.3856,
303
+ "step": 420
304
+ },
305
+ {
306
+ "epoch": 0.30714285714285716,
307
+ "grad_norm": 3.308420181274414,
308
+ "learning_rate": 8.976190476190477e-05,
309
+ "loss": 0.4143,
310
+ "step": 430
311
+ },
312
+ {
313
+ "epoch": 0.3142857142857143,
314
+ "grad_norm": 3.499880075454712,
315
+ "learning_rate": 8.952380952380953e-05,
316
+ "loss": 0.4857,
317
+ "step": 440
318
+ },
319
+ {
320
+ "epoch": 0.32142857142857145,
321
+ "grad_norm": 1.7907614707946777,
322
+ "learning_rate": 8.92857142857143e-05,
323
+ "loss": 0.356,
324
+ "step": 450
325
+ },
326
+ {
327
+ "epoch": 0.32857142857142857,
328
+ "grad_norm": 3.0111501216888428,
329
+ "learning_rate": 8.904761904761905e-05,
330
+ "loss": 0.358,
331
+ "step": 460
332
+ },
333
+ {
334
+ "epoch": 0.3357142857142857,
335
+ "grad_norm": 2.1389689445495605,
336
+ "learning_rate": 8.880952380952381e-05,
337
+ "loss": 0.3331,
338
+ "step": 470
339
+ },
340
+ {
341
+ "epoch": 0.34285714285714286,
342
+ "grad_norm": 1.8007326126098633,
343
+ "learning_rate": 8.857142857142857e-05,
344
+ "loss": 0.3148,
345
+ "step": 480
346
+ },
347
+ {
348
+ "epoch": 0.35,
349
+ "grad_norm": 3.603548288345337,
350
+ "learning_rate": 8.833333333333333e-05,
351
+ "loss": 0.4513,
352
+ "step": 490
353
+ },
354
+ {
355
+ "epoch": 0.35714285714285715,
356
+ "grad_norm": 2.270984411239624,
357
+ "learning_rate": 8.80952380952381e-05,
358
+ "loss": 0.3392,
359
+ "step": 500
360
+ },
361
+ {
362
+ "epoch": 0.36428571428571427,
363
+ "grad_norm": 1.5367858409881592,
364
+ "learning_rate": 8.785714285714286e-05,
365
+ "loss": 0.3665,
366
+ "step": 510
367
+ },
368
+ {
369
+ "epoch": 0.37142857142857144,
370
+ "grad_norm": 1.5397262573242188,
371
+ "learning_rate": 8.761904761904762e-05,
372
+ "loss": 0.3875,
373
+ "step": 520
374
+ },
375
+ {
376
+ "epoch": 0.37857142857142856,
377
+ "grad_norm": 3.6661577224731445,
378
+ "learning_rate": 8.738095238095239e-05,
379
+ "loss": 0.3617,
380
+ "step": 530
381
+ },
382
+ {
383
+ "epoch": 0.38571428571428573,
384
+ "grad_norm": 2.5946245193481445,
385
+ "learning_rate": 8.714285714285715e-05,
386
+ "loss": 0.3318,
387
+ "step": 540
388
+ },
389
+ {
390
+ "epoch": 0.39285714285714285,
391
+ "grad_norm": 2.508056640625,
392
+ "learning_rate": 8.690476190476192e-05,
393
+ "loss": 0.3681,
394
+ "step": 550
395
+ },
396
+ {
397
+ "epoch": 0.4,
398
+ "grad_norm": 5.406072616577148,
399
+ "learning_rate": 8.666666666666667e-05,
400
+ "loss": 0.3589,
401
+ "step": 560
402
+ },
403
+ {
404
+ "epoch": 0.40714285714285714,
405
+ "grad_norm": 2.0363543033599854,
406
+ "learning_rate": 8.642857142857143e-05,
407
+ "loss": 0.4328,
408
+ "step": 570
409
+ },
410
+ {
411
+ "epoch": 0.4142857142857143,
412
+ "grad_norm": 1.796236515045166,
413
+ "learning_rate": 8.61904761904762e-05,
414
+ "loss": 0.4005,
415
+ "step": 580
416
+ },
417
+ {
418
+ "epoch": 0.42142857142857143,
419
+ "grad_norm": 5.6172943115234375,
420
+ "learning_rate": 8.595238095238096e-05,
421
+ "loss": 0.3414,
422
+ "step": 590
423
+ },
424
+ {
425
+ "epoch": 0.42857142857142855,
426
+ "grad_norm": 2.953678846359253,
427
+ "learning_rate": 8.571428571428571e-05,
428
+ "loss": 0.3917,
429
+ "step": 600
430
+ },
431
+ {
432
+ "epoch": 0.4357142857142857,
433
+ "grad_norm": 1.8276687860488892,
434
+ "learning_rate": 8.547619047619048e-05,
435
+ "loss": 0.4057,
436
+ "step": 610
437
+ },
438
+ {
439
+ "epoch": 0.44285714285714284,
440
+ "grad_norm": 1.8668811321258545,
441
+ "learning_rate": 8.523809523809524e-05,
442
+ "loss": 0.4153,
443
+ "step": 620
444
+ },
445
+ {
446
+ "epoch": 0.45,
447
+ "grad_norm": 2.87809419631958,
448
+ "learning_rate": 8.5e-05,
449
+ "loss": 0.3911,
450
+ "step": 630
451
+ },
452
+ {
453
+ "epoch": 0.45714285714285713,
454
+ "grad_norm": 4.863674640655518,
455
+ "learning_rate": 8.476190476190477e-05,
456
+ "loss": 0.4458,
457
+ "step": 640
458
+ },
459
+ {
460
+ "epoch": 0.4642857142857143,
461
+ "grad_norm": 1.7150845527648926,
462
+ "learning_rate": 8.452380952380952e-05,
463
+ "loss": 0.3951,
464
+ "step": 650
465
+ },
466
+ {
467
+ "epoch": 0.4714285714285714,
468
+ "grad_norm": 2.0149145126342773,
469
+ "learning_rate": 8.428571428571429e-05,
470
+ "loss": 0.3238,
471
+ "step": 660
472
+ },
473
+ {
474
+ "epoch": 0.4785714285714286,
475
+ "grad_norm": 2.3146402835845947,
476
+ "learning_rate": 8.404761904761905e-05,
477
+ "loss": 0.3129,
478
+ "step": 670
479
+ },
480
+ {
481
+ "epoch": 0.4857142857142857,
482
+ "grad_norm": 1.8723537921905518,
483
+ "learning_rate": 8.380952380952382e-05,
484
+ "loss": 0.2969,
485
+ "step": 680
486
+ },
487
+ {
488
+ "epoch": 0.4928571428571429,
489
+ "grad_norm": 1.6206679344177246,
490
+ "learning_rate": 8.357142857142858e-05,
491
+ "loss": 0.3673,
492
+ "step": 690
493
+ },
494
+ {
495
+ "epoch": 0.5,
496
+ "grad_norm": 2.490722179412842,
497
+ "learning_rate": 8.333333333333334e-05,
498
+ "loss": 0.3493,
499
+ "step": 700
500
+ },
501
+ {
502
+ "epoch": 0.5071428571428571,
503
+ "grad_norm": 1.7868791818618774,
504
+ "learning_rate": 8.309523809523811e-05,
505
+ "loss": 0.423,
506
+ "step": 710
507
+ },
508
+ {
509
+ "epoch": 0.5142857142857142,
510
+ "grad_norm": 2.3152015209198,
511
+ "learning_rate": 8.285714285714287e-05,
512
+ "loss": 0.3448,
513
+ "step": 720
514
+ },
515
+ {
516
+ "epoch": 0.5214285714285715,
517
+ "grad_norm": 1.934523105621338,
518
+ "learning_rate": 8.261904761904762e-05,
519
+ "loss": 0.372,
520
+ "step": 730
521
+ },
522
+ {
523
+ "epoch": 0.5285714285714286,
524
+ "grad_norm": 2.421624183654785,
525
+ "learning_rate": 8.238095238095238e-05,
526
+ "loss": 0.3602,
527
+ "step": 740
528
+ },
529
+ {
530
+ "epoch": 0.5357142857142857,
531
+ "grad_norm": 2.0156285762786865,
532
+ "learning_rate": 8.214285714285714e-05,
533
+ "loss": 0.334,
534
+ "step": 750
535
+ },
536
+ {
537
+ "epoch": 0.5428571428571428,
538
+ "grad_norm": 1.6513855457305908,
539
+ "learning_rate": 8.19047619047619e-05,
540
+ "loss": 0.3553,
541
+ "step": 760
542
+ },
543
+ {
544
+ "epoch": 0.55,
545
+ "grad_norm": 1.950451374053955,
546
+ "learning_rate": 8.166666666666667e-05,
547
+ "loss": 0.3696,
548
+ "step": 770
549
+ },
550
+ {
551
+ "epoch": 0.5571428571428572,
552
+ "grad_norm": 2.4662423133850098,
553
+ "learning_rate": 8.142857142857143e-05,
554
+ "loss": 0.3232,
555
+ "step": 780
556
+ },
557
+ {
558
+ "epoch": 0.5642857142857143,
559
+ "grad_norm": 2.060654401779175,
560
+ "learning_rate": 8.11904761904762e-05,
561
+ "loss": 0.3294,
562
+ "step": 790
563
+ },
564
+ {
565
+ "epoch": 0.5714285714285714,
566
+ "grad_norm": 3.3581550121307373,
567
+ "learning_rate": 8.095238095238096e-05,
568
+ "loss": 0.3022,
569
+ "step": 800
570
+ },
571
+ {
572
+ "epoch": 0.5785714285714286,
573
+ "grad_norm": 1.6778148412704468,
574
+ "learning_rate": 8.071428571428573e-05,
575
+ "loss": 0.2808,
576
+ "step": 810
577
+ },
578
+ {
579
+ "epoch": 0.5857142857142857,
580
+ "grad_norm": 5.509958267211914,
581
+ "learning_rate": 8.047619047619048e-05,
582
+ "loss": 0.3584,
583
+ "step": 820
584
+ },
585
+ {
586
+ "epoch": 0.5928571428571429,
587
+ "grad_norm": 2.3136510848999023,
588
+ "learning_rate": 8.023809523809524e-05,
589
+ "loss": 0.2967,
590
+ "step": 830
591
+ },
592
+ {
593
+ "epoch": 0.6,
594
+ "grad_norm": 2.757749557495117,
595
+ "learning_rate": 8e-05,
596
+ "loss": 0.3091,
597
+ "step": 840
598
+ },
599
+ {
600
+ "epoch": 0.6071428571428571,
601
+ "grad_norm": 1.6453288793563843,
602
+ "learning_rate": 7.976190476190477e-05,
603
+ "loss": 0.2882,
604
+ "step": 850
605
+ },
606
+ {
607
+ "epoch": 0.6142857142857143,
608
+ "grad_norm": 3.599712371826172,
609
+ "learning_rate": 7.952380952380952e-05,
610
+ "loss": 0.35,
611
+ "step": 860
612
+ },
613
+ {
614
+ "epoch": 0.6214285714285714,
615
+ "grad_norm": 1.203050136566162,
616
+ "learning_rate": 7.928571428571429e-05,
617
+ "loss": 0.3147,
618
+ "step": 870
619
+ },
620
+ {
621
+ "epoch": 0.6285714285714286,
622
+ "grad_norm": 3.0582826137542725,
623
+ "learning_rate": 7.904761904761905e-05,
624
+ "loss": 0.369,
625
+ "step": 880
626
+ },
627
+ {
628
+ "epoch": 0.6357142857142857,
629
+ "grad_norm": 1.7133578062057495,
630
+ "learning_rate": 7.880952380952382e-05,
631
+ "loss": 0.4404,
632
+ "step": 890
633
+ },
634
+ {
635
+ "epoch": 0.6428571428571429,
636
+ "grad_norm": 2.1532819271087646,
637
+ "learning_rate": 7.857142857142858e-05,
638
+ "loss": 0.3249,
639
+ "step": 900
640
+ },
641
+ {
642
+ "epoch": 0.65,
643
+ "grad_norm": 3.529282808303833,
644
+ "learning_rate": 7.833333333333333e-05,
645
+ "loss": 0.3449,
646
+ "step": 910
647
+ },
648
+ {
649
+ "epoch": 0.6571428571428571,
650
+ "grad_norm": 3.412729501724243,
651
+ "learning_rate": 7.80952380952381e-05,
652
+ "loss": 0.3678,
653
+ "step": 920
654
+ },
655
+ {
656
+ "epoch": 0.6642857142857143,
657
+ "grad_norm": 1.8676691055297852,
658
+ "learning_rate": 7.785714285714286e-05,
659
+ "loss": 0.3522,
660
+ "step": 930
661
+ },
662
+ {
663
+ "epoch": 0.6714285714285714,
664
+ "grad_norm": 1.8748834133148193,
665
+ "learning_rate": 7.761904761904762e-05,
666
+ "loss": 0.2561,
667
+ "step": 940
668
+ },
669
+ {
670
+ "epoch": 0.6785714285714286,
671
+ "grad_norm": 3.1622543334960938,
672
+ "learning_rate": 7.738095238095239e-05,
673
+ "loss": 0.3245,
674
+ "step": 950
675
+ },
676
+ {
677
+ "epoch": 0.6857142857142857,
678
+ "grad_norm": 2.17108416557312,
679
+ "learning_rate": 7.714285714285715e-05,
680
+ "loss": 0.335,
681
+ "step": 960
682
+ },
683
+ {
684
+ "epoch": 0.6928571428571428,
685
+ "grad_norm": 3.686673402786255,
686
+ "learning_rate": 7.690476190476192e-05,
687
+ "loss": 0.319,
688
+ "step": 970
689
+ },
690
+ {
691
+ "epoch": 0.7,
692
+ "grad_norm": 7.108774662017822,
693
+ "learning_rate": 7.666666666666667e-05,
694
+ "loss": 0.3337,
695
+ "step": 980
696
+ },
697
+ {
698
+ "epoch": 0.7071428571428572,
699
+ "grad_norm": 2.532336711883545,
700
+ "learning_rate": 7.642857142857143e-05,
701
+ "loss": 0.377,
702
+ "step": 990
703
+ },
704
+ {
705
+ "epoch": 0.7142857142857143,
706
+ "grad_norm": 1.8768038749694824,
707
+ "learning_rate": 7.619047619047618e-05,
708
+ "loss": 0.3745,
709
+ "step": 1000
710
+ },
711
+ {
712
+ "epoch": 0.7214285714285714,
713
+ "grad_norm": 1.8926101922988892,
714
+ "learning_rate": 7.595238095238095e-05,
715
+ "loss": 0.3503,
716
+ "step": 1010
717
+ },
718
+ {
719
+ "epoch": 0.7285714285714285,
720
+ "grad_norm": 2.2763147354125977,
721
+ "learning_rate": 7.571428571428571e-05,
722
+ "loss": 0.307,
723
+ "step": 1020
724
+ },
725
+ {
726
+ "epoch": 0.7357142857142858,
727
+ "grad_norm": 3.663257360458374,
728
+ "learning_rate": 7.547619047619048e-05,
729
+ "loss": 0.2876,
730
+ "step": 1030
731
+ },
732
+ {
733
+ "epoch": 0.7428571428571429,
734
+ "grad_norm": 1.844648003578186,
735
+ "learning_rate": 7.523809523809524e-05,
736
+ "loss": 0.3512,
737
+ "step": 1040
738
+ },
739
+ {
740
+ "epoch": 0.75,
741
+ "grad_norm": 2.5800540447235107,
742
+ "learning_rate": 7.500000000000001e-05,
743
+ "loss": 0.301,
744
+ "step": 1050
745
+ },
746
+ {
747
+ "epoch": 0.7571428571428571,
748
+ "grad_norm": 2.4674232006073,
749
+ "learning_rate": 7.476190476190477e-05,
750
+ "loss": 0.3635,
751
+ "step": 1060
752
+ },
753
+ {
754
+ "epoch": 0.7642857142857142,
755
+ "grad_norm": 3.5119080543518066,
756
+ "learning_rate": 7.452380952380952e-05,
757
+ "loss": 0.3688,
758
+ "step": 1070
759
+ },
760
+ {
761
+ "epoch": 0.7714285714285715,
762
+ "grad_norm": 2.5902345180511475,
763
+ "learning_rate": 7.428571428571429e-05,
764
+ "loss": 0.3136,
765
+ "step": 1080
766
+ },
767
+ {
768
+ "epoch": 0.7785714285714286,
769
+ "grad_norm": 1.734082818031311,
770
+ "learning_rate": 7.404761904761905e-05,
771
+ "loss": 0.3366,
772
+ "step": 1090
773
+ },
774
+ {
775
+ "epoch": 0.7857142857142857,
776
+ "grad_norm": 2.0863406658172607,
777
+ "learning_rate": 7.380952380952382e-05,
778
+ "loss": 0.331,
779
+ "step": 1100
780
+ },
781
+ {
782
+ "epoch": 0.7928571428571428,
783
+ "grad_norm": 2.9227652549743652,
784
+ "learning_rate": 7.357142857142858e-05,
785
+ "loss": 0.3131,
786
+ "step": 1110
787
+ },
788
+ {
789
+ "epoch": 0.8,
790
+ "grad_norm": 5.268605709075928,
791
+ "learning_rate": 7.333333333333333e-05,
792
+ "loss": 0.3212,
793
+ "step": 1120
794
+ },
795
+ {
796
+ "epoch": 0.8071428571428572,
797
+ "grad_norm": 2.633431911468506,
798
+ "learning_rate": 7.30952380952381e-05,
799
+ "loss": 0.2707,
800
+ "step": 1130
801
+ },
802
+ {
803
+ "epoch": 0.8142857142857143,
804
+ "grad_norm": 1.652354121208191,
805
+ "learning_rate": 7.285714285714286e-05,
806
+ "loss": 0.3037,
807
+ "step": 1140
808
+ },
809
+ {
810
+ "epoch": 0.8214285714285714,
811
+ "grad_norm": 3.1987786293029785,
812
+ "learning_rate": 7.261904761904762e-05,
813
+ "loss": 0.3603,
814
+ "step": 1150
815
+ },
816
+ {
817
+ "epoch": 0.8285714285714286,
818
+ "grad_norm": 1.7624555826187134,
819
+ "learning_rate": 7.238095238095238e-05,
820
+ "loss": 0.308,
821
+ "step": 1160
822
+ },
823
+ {
824
+ "epoch": 0.8357142857142857,
825
+ "grad_norm": 2.6379318237304688,
826
+ "learning_rate": 7.214285714285714e-05,
827
+ "loss": 0.3068,
828
+ "step": 1170
829
+ },
830
+ {
831
+ "epoch": 0.8428571428571429,
832
+ "grad_norm": 2.6688101291656494,
833
+ "learning_rate": 7.19047619047619e-05,
834
+ "loss": 0.3134,
835
+ "step": 1180
836
+ },
837
+ {
838
+ "epoch": 0.85,
839
+ "grad_norm": 3.698643207550049,
840
+ "learning_rate": 7.166666666666667e-05,
841
+ "loss": 0.2973,
842
+ "step": 1190
843
+ },
844
+ {
845
+ "epoch": 0.8571428571428571,
846
+ "grad_norm": 2.9004178047180176,
847
+ "learning_rate": 7.142857142857143e-05,
848
+ "loss": 0.3116,
849
+ "step": 1200
850
+ },
851
+ {
852
+ "epoch": 0.8642857142857143,
853
+ "grad_norm": 2.3414883613586426,
854
+ "learning_rate": 7.11904761904762e-05,
855
+ "loss": 0.334,
856
+ "step": 1210
857
+ },
858
+ {
859
+ "epoch": 0.8714285714285714,
860
+ "grad_norm": 3.6203651428222656,
861
+ "learning_rate": 7.095238095238096e-05,
862
+ "loss": 0.3241,
863
+ "step": 1220
864
+ },
865
+ {
866
+ "epoch": 0.8785714285714286,
867
+ "grad_norm": 2.7472829818725586,
868
+ "learning_rate": 7.071428571428573e-05,
869
+ "loss": 0.2727,
870
+ "step": 1230
871
+ },
872
+ {
873
+ "epoch": 0.8857142857142857,
874
+ "grad_norm": 3.2332115173339844,
875
+ "learning_rate": 7.047619047619048e-05,
876
+ "loss": 0.3329,
877
+ "step": 1240
878
+ },
879
+ {
880
+ "epoch": 0.8928571428571429,
881
+ "grad_norm": 1.968595027923584,
882
+ "learning_rate": 7.023809523809524e-05,
883
+ "loss": 0.3478,
884
+ "step": 1250
885
+ },
886
+ {
887
+ "epoch": 0.9,
888
+ "grad_norm": 2.3712966442108154,
889
+ "learning_rate": 7e-05,
890
+ "loss": 0.3964,
891
+ "step": 1260
892
+ },
893
+ {
894
+ "epoch": 0.9071428571428571,
895
+ "grad_norm": 1.7704048156738281,
896
+ "learning_rate": 6.976190476190476e-05,
897
+ "loss": 0.3983,
898
+ "step": 1270
899
+ },
900
+ {
901
+ "epoch": 0.9142857142857143,
902
+ "grad_norm": 1.8879201412200928,
903
+ "learning_rate": 6.952380952380952e-05,
904
+ "loss": 0.3296,
905
+ "step": 1280
906
+ },
907
+ {
908
+ "epoch": 0.9214285714285714,
909
+ "grad_norm": 4.0522780418396,
910
+ "learning_rate": 6.928571428571429e-05,
911
+ "loss": 0.3722,
912
+ "step": 1290
913
+ },
914
+ {
915
+ "epoch": 0.9285714285714286,
916
+ "grad_norm": 1.7849159240722656,
917
+ "learning_rate": 6.904761904761905e-05,
918
+ "loss": 0.3022,
919
+ "step": 1300
920
+ },
921
+ {
922
+ "epoch": 0.9357142857142857,
923
+ "grad_norm": 3.376678705215454,
924
+ "learning_rate": 6.880952380952382e-05,
925
+ "loss": 0.3429,
926
+ "step": 1310
927
+ },
928
+ {
929
+ "epoch": 0.9428571428571428,
930
+ "grad_norm": 4.100617408752441,
931
+ "learning_rate": 6.857142857142858e-05,
932
+ "loss": 0.2793,
933
+ "step": 1320
934
+ },
935
+ {
936
+ "epoch": 0.95,
937
+ "grad_norm": 3.8000288009643555,
938
+ "learning_rate": 6.833333333333333e-05,
939
+ "loss": 0.3114,
940
+ "step": 1330
941
+ },
942
+ {
943
+ "epoch": 0.9571428571428572,
944
+ "grad_norm": 4.846517562866211,
945
+ "learning_rate": 6.80952380952381e-05,
946
+ "loss": 0.3377,
947
+ "step": 1340
948
+ },
949
+ {
950
+ "epoch": 0.9642857142857143,
951
+ "grad_norm": 2.0567409992218018,
952
+ "learning_rate": 6.785714285714286e-05,
953
+ "loss": 0.3198,
954
+ "step": 1350
955
+ },
956
+ {
957
+ "epoch": 0.9714285714285714,
958
+ "grad_norm": 2.3197948932647705,
959
+ "learning_rate": 6.761904761904763e-05,
960
+ "loss": 0.2817,
961
+ "step": 1360
962
+ },
963
+ {
964
+ "epoch": 0.9785714285714285,
965
+ "grad_norm": 2.7312827110290527,
966
+ "learning_rate": 6.738095238095239e-05,
967
+ "loss": 0.3378,
968
+ "step": 1370
969
+ },
970
+ {
971
+ "epoch": 0.9857142857142858,
972
+ "grad_norm": 3.115516185760498,
973
+ "learning_rate": 6.714285714285714e-05,
974
+ "loss": 0.3638,
975
+ "step": 1380
976
+ },
977
+ {
978
+ "epoch": 0.9928571428571429,
979
+ "grad_norm": 1.0742294788360596,
980
+ "learning_rate": 6.69047619047619e-05,
981
+ "loss": 0.3176,
982
+ "step": 1390
983
+ },
984
+ {
985
+ "epoch": 1.0,
986
+ "grad_norm": 2.3872761726379395,
987
+ "learning_rate": 6.666666666666667e-05,
988
+ "loss": 0.2921,
989
+ "step": 1400
990
+ },
991
+ {
992
+ "epoch": 1.0,
993
+ "eval_loss": 0.2929375469684601,
994
+ "eval_runtime": 68.5646,
995
+ "eval_samples_per_second": 2.917,
996
+ "eval_steps_per_second": 0.365,
997
+ "step": 1400
998
+ },
999
+ {
1000
+ "epoch": 1.0071428571428571,
1001
+ "grad_norm": 2.1662676334381104,
1002
+ "learning_rate": 6.642857142857143e-05,
1003
+ "loss": 0.2885,
1004
+ "step": 1410
1005
+ },
1006
+ {
1007
+ "epoch": 1.0142857142857142,
1008
+ "grad_norm": 4.116220474243164,
1009
+ "learning_rate": 6.619047619047619e-05,
1010
+ "loss": 0.2951,
1011
+ "step": 1420
1012
+ },
1013
+ {
1014
+ "epoch": 1.0214285714285714,
1015
+ "grad_norm": 4.550313472747803,
1016
+ "learning_rate": 6.595238095238095e-05,
1017
+ "loss": 0.3078,
1018
+ "step": 1430
1019
+ },
1020
+ {
1021
+ "epoch": 1.0285714285714285,
1022
+ "grad_norm": 1.7829937934875488,
1023
+ "learning_rate": 6.571428571428571e-05,
1024
+ "loss": 0.3309,
1025
+ "step": 1440
1026
+ },
1027
+ {
1028
+ "epoch": 1.0357142857142858,
1029
+ "grad_norm": 2.684500217437744,
1030
+ "learning_rate": 6.547619047619048e-05,
1031
+ "loss": 0.2994,
1032
+ "step": 1450
1033
+ },
1034
+ {
1035
+ "epoch": 1.042857142857143,
1036
+ "grad_norm": 1.3234150409698486,
1037
+ "learning_rate": 6.523809523809524e-05,
1038
+ "loss": 0.3782,
1039
+ "step": 1460
1040
+ },
1041
+ {
1042
+ "epoch": 1.05,
1043
+ "grad_norm": 2.6253223419189453,
1044
+ "learning_rate": 6.500000000000001e-05,
1045
+ "loss": 0.3094,
1046
+ "step": 1470
1047
+ },
1048
+ {
1049
+ "epoch": 1.0571428571428572,
1050
+ "grad_norm": 3.272584915161133,
1051
+ "learning_rate": 6.476190476190477e-05,
1052
+ "loss": 0.3406,
1053
+ "step": 1480
1054
+ },
1055
+ {
1056
+ "epoch": 1.0642857142857143,
1057
+ "grad_norm": 3.765272378921509,
1058
+ "learning_rate": 6.452380952380954e-05,
1059
+ "loss": 0.3525,
1060
+ "step": 1490
1061
+ },
1062
+ {
1063
+ "epoch": 1.0714285714285714,
1064
+ "grad_norm": 4.5847578048706055,
1065
+ "learning_rate": 6.428571428571429e-05,
1066
+ "loss": 0.3296,
1067
+ "step": 1500
1068
+ },
1069
+ {
1070
+ "epoch": 1.0785714285714285,
1071
+ "grad_norm": 2.448350191116333,
1072
+ "learning_rate": 6.404761904761904e-05,
1073
+ "loss": 0.2704,
1074
+ "step": 1510
1075
+ },
1076
+ {
1077
+ "epoch": 1.0857142857142856,
1078
+ "grad_norm": 2.5641868114471436,
1079
+ "learning_rate": 6.38095238095238e-05,
1080
+ "loss": 0.277,
1081
+ "step": 1520
1082
+ },
1083
+ {
1084
+ "epoch": 1.092857142857143,
1085
+ "grad_norm": 1.8631259202957153,
1086
+ "learning_rate": 6.357142857142857e-05,
1087
+ "loss": 0.2915,
1088
+ "step": 1530
1089
+ },
1090
+ {
1091
+ "epoch": 1.1,
1092
+ "grad_norm": 1.7823829650878906,
1093
+ "learning_rate": 6.333333333333333e-05,
1094
+ "loss": 0.279,
1095
+ "step": 1540
1096
+ },
1097
+ {
1098
+ "epoch": 1.1071428571428572,
1099
+ "grad_norm": 3.2730162143707275,
1100
+ "learning_rate": 6.30952380952381e-05,
1101
+ "loss": 0.3179,
1102
+ "step": 1550
1103
+ },
1104
+ {
1105
+ "epoch": 1.1142857142857143,
1106
+ "grad_norm": 2.495410203933716,
1107
+ "learning_rate": 6.285714285714286e-05,
1108
+ "loss": 0.326,
1109
+ "step": 1560
1110
+ },
1111
+ {
1112
+ "epoch": 1.1214285714285714,
1113
+ "grad_norm": 1.844226598739624,
1114
+ "learning_rate": 6.261904761904763e-05,
1115
+ "loss": 0.2931,
1116
+ "step": 1570
1117
+ },
1118
+ {
1119
+ "epoch": 1.1285714285714286,
1120
+ "grad_norm": 3.5361154079437256,
1121
+ "learning_rate": 6.238095238095239e-05,
1122
+ "loss": 0.2932,
1123
+ "step": 1580
1124
+ },
1125
+ {
1126
+ "epoch": 1.1357142857142857,
1127
+ "grad_norm": 1.9047455787658691,
1128
+ "learning_rate": 6.214285714285714e-05,
1129
+ "loss": 0.2779,
1130
+ "step": 1590
1131
+ },
1132
+ {
1133
+ "epoch": 1.1428571428571428,
1134
+ "grad_norm": 6.630354881286621,
1135
+ "learning_rate": 6.19047619047619e-05,
1136
+ "loss": 0.3563,
1137
+ "step": 1600
1138
+ },
1139
+ {
1140
+ "epoch": 1.15,
1141
+ "grad_norm": 4.267199993133545,
1142
+ "learning_rate": 6.166666666666667e-05,
1143
+ "loss": 0.3146,
1144
+ "step": 1610
1145
+ },
1146
+ {
1147
+ "epoch": 1.157142857142857,
1148
+ "grad_norm": 4.094973087310791,
1149
+ "learning_rate": 6.142857142857143e-05,
1150
+ "loss": 0.3096,
1151
+ "step": 1620
1152
+ },
1153
+ {
1154
+ "epoch": 1.1642857142857144,
1155
+ "grad_norm": 2.1625826358795166,
1156
+ "learning_rate": 6.11904761904762e-05,
1157
+ "loss": 0.2893,
1158
+ "step": 1630
1159
+ },
1160
+ {
1161
+ "epoch": 1.1714285714285715,
1162
+ "grad_norm": 4.5134758949279785,
1163
+ "learning_rate": 6.0952380952380964e-05,
1164
+ "loss": 0.3037,
1165
+ "step": 1640
1166
+ },
1167
+ {
1168
+ "epoch": 1.1785714285714286,
1169
+ "grad_norm": 1.987043023109436,
1170
+ "learning_rate": 6.0714285714285715e-05,
1171
+ "loss": 0.3427,
1172
+ "step": 1650
1173
+ },
1174
+ {
1175
+ "epoch": 1.1857142857142857,
1176
+ "grad_norm": 1.7972902059555054,
1177
+ "learning_rate": 6.047619047619047e-05,
1178
+ "loss": 0.3012,
1179
+ "step": 1660
1180
+ },
1181
+ {
1182
+ "epoch": 1.1928571428571428,
1183
+ "grad_norm": 2.2862155437469482,
1184
+ "learning_rate": 6.023809523809524e-05,
1185
+ "loss": 0.2877,
1186
+ "step": 1670
1187
+ },
1188
+ {
1189
+ "epoch": 1.2,
1190
+ "grad_norm": 2.5027403831481934,
1191
+ "learning_rate": 6e-05,
1192
+ "loss": 0.3172,
1193
+ "step": 1680
1194
+ },
1195
+ {
1196
+ "epoch": 1.207142857142857,
1197
+ "grad_norm": 2.698021173477173,
1198
+ "learning_rate": 5.9761904761904766e-05,
1199
+ "loss": 0.2689,
1200
+ "step": 1690
1201
+ },
1202
+ {
1203
+ "epoch": 1.2142857142857142,
1204
+ "grad_norm": 1.9757822751998901,
1205
+ "learning_rate": 5.9523809523809524e-05,
1206
+ "loss": 0.2845,
1207
+ "step": 1700
1208
+ },
1209
+ {
1210
+ "epoch": 1.2214285714285715,
1211
+ "grad_norm": 1.7587430477142334,
1212
+ "learning_rate": 5.928571428571429e-05,
1213
+ "loss": 0.3264,
1214
+ "step": 1710
1215
+ },
1216
+ {
1217
+ "epoch": 1.2285714285714286,
1218
+ "grad_norm": 2.466726541519165,
1219
+ "learning_rate": 5.904761904761905e-05,
1220
+ "loss": 0.2311,
1221
+ "step": 1720
1222
+ },
1223
+ {
1224
+ "epoch": 1.2357142857142858,
1225
+ "grad_norm": 2.041752815246582,
1226
+ "learning_rate": 5.880952380952382e-05,
1227
+ "loss": 0.2829,
1228
+ "step": 1730
1229
+ },
1230
+ {
1231
+ "epoch": 1.2428571428571429,
1232
+ "grad_norm": 4.565363883972168,
1233
+ "learning_rate": 5.8571428571428575e-05,
1234
+ "loss": 0.2356,
1235
+ "step": 1740
1236
+ },
1237
+ {
1238
+ "epoch": 1.25,
1239
+ "grad_norm": 3.1162221431732178,
1240
+ "learning_rate": 5.833333333333334e-05,
1241
+ "loss": 0.2719,
1242
+ "step": 1750
1243
+ },
1244
+ {
1245
+ "epoch": 1.2571428571428571,
1246
+ "grad_norm": 4.6578779220581055,
1247
+ "learning_rate": 5.8095238095238104e-05,
1248
+ "loss": 0.2477,
1249
+ "step": 1760
1250
+ },
1251
+ {
1252
+ "epoch": 1.2642857142857142,
1253
+ "grad_norm": 2.782367706298828,
1254
+ "learning_rate": 5.785714285714287e-05,
1255
+ "loss": 0.3718,
1256
+ "step": 1770
1257
+ },
1258
+ {
1259
+ "epoch": 1.2714285714285714,
1260
+ "grad_norm": 3.4065630435943604,
1261
+ "learning_rate": 5.761904761904762e-05,
1262
+ "loss": 0.3211,
1263
+ "step": 1780
1264
+ },
1265
+ {
1266
+ "epoch": 1.2785714285714285,
1267
+ "grad_norm": 2.6382477283477783,
1268
+ "learning_rate": 5.738095238095238e-05,
1269
+ "loss": 0.3476,
1270
+ "step": 1790
1271
+ },
1272
+ {
1273
+ "epoch": 1.2857142857142856,
1274
+ "grad_norm": 1.6404658555984497,
1275
+ "learning_rate": 5.714285714285714e-05,
1276
+ "loss": 0.3465,
1277
+ "step": 1800
1278
+ },
1279
+ {
1280
+ "epoch": 1.292857142857143,
1281
+ "grad_norm": 2.675689935684204,
1282
+ "learning_rate": 5.6904761904761906e-05,
1283
+ "loss": 0.3027,
1284
+ "step": 1810
1285
+ },
1286
+ {
1287
+ "epoch": 1.3,
1288
+ "grad_norm": 3.062359571456909,
1289
+ "learning_rate": 5.666666666666667e-05,
1290
+ "loss": 0.3043,
1291
+ "step": 1820
1292
+ },
1293
+ {
1294
+ "epoch": 1.3071428571428572,
1295
+ "grad_norm": 3.5478060245513916,
1296
+ "learning_rate": 5.642857142857143e-05,
1297
+ "loss": 0.3022,
1298
+ "step": 1830
1299
+ },
1300
+ {
1301
+ "epoch": 1.3142857142857143,
1302
+ "grad_norm": 3.2227487564086914,
1303
+ "learning_rate": 5.619047619047619e-05,
1304
+ "loss": 0.2776,
1305
+ "step": 1840
1306
+ },
1307
+ {
1308
+ "epoch": 1.3214285714285714,
1309
+ "grad_norm": 1.7066378593444824,
1310
+ "learning_rate": 5.595238095238096e-05,
1311
+ "loss": 0.3345,
1312
+ "step": 1850
1313
+ },
1314
+ {
1315
+ "epoch": 1.3285714285714285,
1316
+ "grad_norm": 3.7412548065185547,
1317
+ "learning_rate": 5.571428571428572e-05,
1318
+ "loss": 0.2886,
1319
+ "step": 1860
1320
+ },
1321
+ {
1322
+ "epoch": 1.3357142857142856,
1323
+ "grad_norm": 6.295001983642578,
1324
+ "learning_rate": 5.547619047619048e-05,
1325
+ "loss": 0.2926,
1326
+ "step": 1870
1327
+ },
1328
+ {
1329
+ "epoch": 1.342857142857143,
1330
+ "grad_norm": 2.9826364517211914,
1331
+ "learning_rate": 5.5238095238095244e-05,
1332
+ "loss": 0.2908,
1333
+ "step": 1880
1334
+ },
1335
+ {
1336
+ "epoch": 1.35,
1337
+ "grad_norm": 2.300792932510376,
1338
+ "learning_rate": 5.500000000000001e-05,
1339
+ "loss": 0.3346,
1340
+ "step": 1890
1341
+ },
1342
+ {
1343
+ "epoch": 1.3571428571428572,
1344
+ "grad_norm": 3.02064847946167,
1345
+ "learning_rate": 5.4761904761904766e-05,
1346
+ "loss": 0.298,
1347
+ "step": 1900
1348
+ },
1349
+ {
1350
+ "epoch": 1.3642857142857143,
1351
+ "grad_norm": 2.3579671382904053,
1352
+ "learning_rate": 5.4523809523809524e-05,
1353
+ "loss": 0.3148,
1354
+ "step": 1910
1355
+ },
1356
+ {
1357
+ "epoch": 1.3714285714285714,
1358
+ "grad_norm": 2.4837729930877686,
1359
+ "learning_rate": 5.428571428571428e-05,
1360
+ "loss": 0.2942,
1361
+ "step": 1920
1362
+ },
1363
+ {
1364
+ "epoch": 1.3785714285714286,
1365
+ "grad_norm": 1.5465370416641235,
1366
+ "learning_rate": 5.4047619047619046e-05,
1367
+ "loss": 0.3141,
1368
+ "step": 1930
1369
+ },
1370
+ {
1371
+ "epoch": 1.3857142857142857,
1372
+ "grad_norm": 3.10286808013916,
1373
+ "learning_rate": 5.380952380952381e-05,
1374
+ "loss": 0.2519,
1375
+ "step": 1940
1376
+ },
1377
+ {
1378
+ "epoch": 1.3928571428571428,
1379
+ "grad_norm": 2.024423837661743,
1380
+ "learning_rate": 5.3571428571428575e-05,
1381
+ "loss": 0.2399,
1382
+ "step": 1950
1383
+ },
1384
+ {
1385
+ "epoch": 1.4,
1386
+ "grad_norm": 3.297781467437744,
1387
+ "learning_rate": 5.333333333333333e-05,
1388
+ "loss": 0.3115,
1389
+ "step": 1960
1390
+ },
1391
+ {
1392
+ "epoch": 1.407142857142857,
1393
+ "grad_norm": 12.920825004577637,
1394
+ "learning_rate": 5.30952380952381e-05,
1395
+ "loss": 0.3199,
1396
+ "step": 1970
1397
+ },
1398
+ {
1399
+ "epoch": 1.4142857142857144,
1400
+ "grad_norm": 1.6162301301956177,
1401
+ "learning_rate": 5.285714285714286e-05,
1402
+ "loss": 0.228,
1403
+ "step": 1980
1404
+ },
1405
+ {
1406
+ "epoch": 1.4214285714285715,
1407
+ "grad_norm": 3.950810194015503,
1408
+ "learning_rate": 5.261904761904763e-05,
1409
+ "loss": 0.3363,
1410
+ "step": 1990
1411
+ },
1412
+ {
1413
+ "epoch": 1.4285714285714286,
1414
+ "grad_norm": 1.5367783308029175,
1415
+ "learning_rate": 5.2380952380952384e-05,
1416
+ "loss": 0.2328,
1417
+ "step": 2000
1418
+ },
1419
+ {
1420
+ "epoch": 1.4357142857142857,
1421
+ "grad_norm": 1.4729127883911133,
1422
+ "learning_rate": 5.214285714285715e-05,
1423
+ "loss": 0.2892,
1424
+ "step": 2010
1425
+ },
1426
+ {
1427
+ "epoch": 1.4428571428571428,
1428
+ "grad_norm": 3.7027993202209473,
1429
+ "learning_rate": 5.1904761904761913e-05,
1430
+ "loss": 0.2598,
1431
+ "step": 2020
1432
+ },
1433
+ {
1434
+ "epoch": 1.45,
1435
+ "grad_norm": 4.046818256378174,
1436
+ "learning_rate": 5.166666666666667e-05,
1437
+ "loss": 0.3184,
1438
+ "step": 2030
1439
+ },
1440
+ {
1441
+ "epoch": 1.457142857142857,
1442
+ "grad_norm": 5.93725061416626,
1443
+ "learning_rate": 5.142857142857143e-05,
1444
+ "loss": 0.2777,
1445
+ "step": 2040
1446
+ },
1447
+ {
1448
+ "epoch": 1.4642857142857144,
1449
+ "grad_norm": 2.02740740776062,
1450
+ "learning_rate": 5.119047619047619e-05,
1451
+ "loss": 0.2632,
1452
+ "step": 2050
1453
+ },
1454
+ {
1455
+ "epoch": 1.4714285714285715,
1456
+ "grad_norm": 1.166384220123291,
1457
+ "learning_rate": 5.095238095238095e-05,
1458
+ "loss": 0.2457,
1459
+ "step": 2060
1460
+ },
1461
+ {
1462
+ "epoch": 1.4785714285714286,
1463
+ "grad_norm": 2.251250982284546,
1464
+ "learning_rate": 5.0714285714285716e-05,
1465
+ "loss": 0.2996,
1466
+ "step": 2070
1467
+ },
1468
+ {
1469
+ "epoch": 1.4857142857142858,
1470
+ "grad_norm": 1.698370337486267,
1471
+ "learning_rate": 5.047619047619048e-05,
1472
+ "loss": 0.2418,
1473
+ "step": 2080
1474
+ },
1475
+ {
1476
+ "epoch": 1.4928571428571429,
1477
+ "grad_norm": 1.665677785873413,
1478
+ "learning_rate": 5.023809523809524e-05,
1479
+ "loss": 0.2718,
1480
+ "step": 2090
1481
+ },
1482
+ {
1483
+ "epoch": 1.5,
1484
+ "grad_norm": 2.417782783508301,
1485
+ "learning_rate": 5e-05,
1486
+ "loss": 0.263,
1487
+ "step": 2100
1488
+ },
1489
+ {
1490
+ "epoch": 1.5071428571428571,
1491
+ "grad_norm": 3.639488935470581,
1492
+ "learning_rate": 4.976190476190477e-05,
1493
+ "loss": 0.2855,
1494
+ "step": 2110
1495
+ },
1496
+ {
1497
+ "epoch": 1.5142857142857142,
1498
+ "grad_norm": 3.0365843772888184,
1499
+ "learning_rate": 4.9523809523809525e-05,
1500
+ "loss": 0.3073,
1501
+ "step": 2120
1502
+ },
1503
+ {
1504
+ "epoch": 1.5214285714285714,
1505
+ "grad_norm": 4.151659965515137,
1506
+ "learning_rate": 4.928571428571429e-05,
1507
+ "loss": 0.2918,
1508
+ "step": 2130
1509
+ },
1510
+ {
1511
+ "epoch": 1.5285714285714285,
1512
+ "grad_norm": 11.354039192199707,
1513
+ "learning_rate": 4.904761904761905e-05,
1514
+ "loss": 0.3428,
1515
+ "step": 2140
1516
+ },
1517
+ {
1518
+ "epoch": 1.5357142857142856,
1519
+ "grad_norm": 3.684797525405884,
1520
+ "learning_rate": 4.880952380952381e-05,
1521
+ "loss": 0.2791,
1522
+ "step": 2150
1523
+ },
1524
+ {
1525
+ "epoch": 1.5428571428571427,
1526
+ "grad_norm": 3.1418874263763428,
1527
+ "learning_rate": 4.8571428571428576e-05,
1528
+ "loss": 0.2386,
1529
+ "step": 2160
1530
+ },
1531
+ {
1532
+ "epoch": 1.55,
1533
+ "grad_norm": 2.303208827972412,
1534
+ "learning_rate": 4.8333333333333334e-05,
1535
+ "loss": 0.3338,
1536
+ "step": 2170
1537
+ },
1538
+ {
1539
+ "epoch": 1.5571428571428572,
1540
+ "grad_norm": 7.349155902862549,
1541
+ "learning_rate": 4.80952380952381e-05,
1542
+ "loss": 0.2595,
1543
+ "step": 2180
1544
+ },
1545
+ {
1546
+ "epoch": 1.5642857142857143,
1547
+ "grad_norm": 2.4980363845825195,
1548
+ "learning_rate": 4.785714285714286e-05,
1549
+ "loss": 0.2499,
1550
+ "step": 2190
1551
+ },
1552
+ {
1553
+ "epoch": 1.5714285714285714,
1554
+ "grad_norm": 3.0202884674072266,
1555
+ "learning_rate": 4.761904761904762e-05,
1556
+ "loss": 0.2608,
1557
+ "step": 2200
1558
+ },
1559
+ {
1560
+ "epoch": 1.5785714285714287,
1561
+ "grad_norm": 7.224872589111328,
1562
+ "learning_rate": 4.738095238095238e-05,
1563
+ "loss": 0.3299,
1564
+ "step": 2210
1565
+ },
1566
+ {
1567
+ "epoch": 1.5857142857142859,
1568
+ "grad_norm": 3.9374120235443115,
1569
+ "learning_rate": 4.714285714285714e-05,
1570
+ "loss": 0.3279,
1571
+ "step": 2220
1572
+ },
1573
+ {
1574
+ "epoch": 1.592857142857143,
1575
+ "grad_norm": 5.184910774230957,
1576
+ "learning_rate": 4.690476190476191e-05,
1577
+ "loss": 0.2691,
1578
+ "step": 2230
1579
+ },
1580
+ {
1581
+ "epoch": 1.6,
1582
+ "grad_norm": 5.877132415771484,
1583
+ "learning_rate": 4.666666666666667e-05,
1584
+ "loss": 0.2903,
1585
+ "step": 2240
1586
+ },
1587
+ {
1588
+ "epoch": 1.6071428571428572,
1589
+ "grad_norm": 3.351426601409912,
1590
+ "learning_rate": 4.642857142857143e-05,
1591
+ "loss": 0.2771,
1592
+ "step": 2250
1593
+ },
1594
+ {
1595
+ "epoch": 1.6142857142857143,
1596
+ "grad_norm": 1.7859790325164795,
1597
+ "learning_rate": 4.6190476190476194e-05,
1598
+ "loss": 0.349,
1599
+ "step": 2260
1600
+ },
1601
+ {
1602
+ "epoch": 1.6214285714285714,
1603
+ "grad_norm": 1.9868186712265015,
1604
+ "learning_rate": 4.595238095238095e-05,
1605
+ "loss": 0.3505,
1606
+ "step": 2270
1607
+ },
1608
+ {
1609
+ "epoch": 1.6285714285714286,
1610
+ "grad_norm": 3.075303316116333,
1611
+ "learning_rate": 4.5714285714285716e-05,
1612
+ "loss": 0.2753,
1613
+ "step": 2280
1614
+ },
1615
+ {
1616
+ "epoch": 1.6357142857142857,
1617
+ "grad_norm": 2.1696481704711914,
1618
+ "learning_rate": 4.547619047619048e-05,
1619
+ "loss": 0.2501,
1620
+ "step": 2290
1621
+ },
1622
+ {
1623
+ "epoch": 1.6428571428571428,
1624
+ "grad_norm": 2.9309706687927246,
1625
+ "learning_rate": 4.523809523809524e-05,
1626
+ "loss": 0.2904,
1627
+ "step": 2300
1628
+ },
1629
+ {
1630
+ "epoch": 1.65,
1631
+ "grad_norm": 3.5121254920959473,
1632
+ "learning_rate": 4.5e-05,
1633
+ "loss": 0.2786,
1634
+ "step": 2310
1635
+ },
1636
+ {
1637
+ "epoch": 1.657142857142857,
1638
+ "grad_norm": 1.3729180097579956,
1639
+ "learning_rate": 4.476190476190477e-05,
1640
+ "loss": 0.2589,
1641
+ "step": 2320
1642
+ },
1643
+ {
1644
+ "epoch": 1.6642857142857141,
1645
+ "grad_norm": 4.851173400878906,
1646
+ "learning_rate": 4.4523809523809525e-05,
1647
+ "loss": 0.3489,
1648
+ "step": 2330
1649
+ },
1650
+ {
1651
+ "epoch": 1.6714285714285713,
1652
+ "grad_norm": 4.376727104187012,
1653
+ "learning_rate": 4.428571428571428e-05,
1654
+ "loss": 0.3188,
1655
+ "step": 2340
1656
+ },
1657
+ {
1658
+ "epoch": 1.6785714285714286,
1659
+ "grad_norm": 2.563443183898926,
1660
+ "learning_rate": 4.404761904761905e-05,
1661
+ "loss": 0.3499,
1662
+ "step": 2350
1663
+ },
1664
+ {
1665
+ "epoch": 1.6857142857142857,
1666
+ "grad_norm": 2.768203020095825,
1667
+ "learning_rate": 4.380952380952381e-05,
1668
+ "loss": 0.2616,
1669
+ "step": 2360
1670
+ },
1671
+ {
1672
+ "epoch": 1.6928571428571428,
1673
+ "grad_norm": 4.015028476715088,
1674
+ "learning_rate": 4.3571428571428576e-05,
1675
+ "loss": 0.2851,
1676
+ "step": 2370
1677
+ },
1678
+ {
1679
+ "epoch": 1.7,
1680
+ "grad_norm": 1.4554102420806885,
1681
+ "learning_rate": 4.3333333333333334e-05,
1682
+ "loss": 0.274,
1683
+ "step": 2380
1684
+ },
1685
+ {
1686
+ "epoch": 1.7071428571428573,
1687
+ "grad_norm": 2.553905963897705,
1688
+ "learning_rate": 4.30952380952381e-05,
1689
+ "loss": 0.2692,
1690
+ "step": 2390
1691
+ },
1692
+ {
1693
+ "epoch": 1.7142857142857144,
1694
+ "grad_norm": 3.319626569747925,
1695
+ "learning_rate": 4.2857142857142856e-05,
1696
+ "loss": 0.3143,
1697
+ "step": 2400
1698
+ },
1699
+ {
1700
+ "epoch": 1.7214285714285715,
1701
+ "grad_norm": 2.207327127456665,
1702
+ "learning_rate": 4.261904761904762e-05,
1703
+ "loss": 0.2548,
1704
+ "step": 2410
1705
+ },
1706
+ {
1707
+ "epoch": 1.7285714285714286,
1708
+ "grad_norm": 3.1698238849639893,
1709
+ "learning_rate": 4.2380952380952385e-05,
1710
+ "loss": 0.3354,
1711
+ "step": 2420
1712
+ },
1713
+ {
1714
+ "epoch": 1.7357142857142858,
1715
+ "grad_norm": 5.368778228759766,
1716
+ "learning_rate": 4.214285714285714e-05,
1717
+ "loss": 0.2665,
1718
+ "step": 2430
1719
+ },
1720
+ {
1721
+ "epoch": 1.7428571428571429,
1722
+ "grad_norm": 1.1383056640625,
1723
+ "learning_rate": 4.190476190476191e-05,
1724
+ "loss": 0.2947,
1725
+ "step": 2440
1726
+ },
1727
+ {
1728
+ "epoch": 1.75,
1729
+ "grad_norm": 3.5954413414001465,
1730
+ "learning_rate": 4.166666666666667e-05,
1731
+ "loss": 0.3362,
1732
+ "step": 2450
1733
+ },
1734
+ {
1735
+ "epoch": 1.7571428571428571,
1736
+ "grad_norm": 3.8164236545562744,
1737
+ "learning_rate": 4.1428571428571437e-05,
1738
+ "loss": 0.2375,
1739
+ "step": 2460
1740
+ },
1741
+ {
1742
+ "epoch": 1.7642857142857142,
1743
+ "grad_norm": 4.327922344207764,
1744
+ "learning_rate": 4.119047619047619e-05,
1745
+ "loss": 0.2776,
1746
+ "step": 2470
1747
+ },
1748
+ {
1749
+ "epoch": 1.7714285714285714,
1750
+ "grad_norm": 2.2782583236694336,
1751
+ "learning_rate": 4.095238095238095e-05,
1752
+ "loss": 0.2407,
1753
+ "step": 2480
1754
+ },
1755
+ {
1756
+ "epoch": 1.7785714285714285,
1757
+ "grad_norm": 1.7074918746948242,
1758
+ "learning_rate": 4.0714285714285717e-05,
1759
+ "loss": 0.2514,
1760
+ "step": 2490
1761
+ },
1762
+ {
1763
+ "epoch": 1.7857142857142856,
1764
+ "grad_norm": 4.881974220275879,
1765
+ "learning_rate": 4.047619047619048e-05,
1766
+ "loss": 0.3229,
1767
+ "step": 2500
1768
+ },
1769
+ {
1770
+ "epoch": 1.7928571428571427,
1771
+ "grad_norm": 4.011612892150879,
1772
+ "learning_rate": 4.023809523809524e-05,
1773
+ "loss": 0.3093,
1774
+ "step": 2510
1775
+ },
1776
+ {
1777
+ "epoch": 1.8,
1778
+ "grad_norm": 2.8738279342651367,
1779
+ "learning_rate": 4e-05,
1780
+ "loss": 0.2844,
1781
+ "step": 2520
1782
+ },
1783
+ {
1784
+ "epoch": 1.8071428571428572,
1785
+ "grad_norm": 2.006197214126587,
1786
+ "learning_rate": 3.976190476190476e-05,
1787
+ "loss": 0.2299,
1788
+ "step": 2530
1789
+ },
1790
+ {
1791
+ "epoch": 1.8142857142857143,
1792
+ "grad_norm": 2.935206651687622,
1793
+ "learning_rate": 3.9523809523809526e-05,
1794
+ "loss": 0.2737,
1795
+ "step": 2540
1796
+ },
1797
+ {
1798
+ "epoch": 1.8214285714285714,
1799
+ "grad_norm": 2.227921962738037,
1800
+ "learning_rate": 3.928571428571429e-05,
1801
+ "loss": 0.2654,
1802
+ "step": 2550
1803
+ },
1804
+ {
1805
+ "epoch": 1.8285714285714287,
1806
+ "grad_norm": 5.1243367195129395,
1807
+ "learning_rate": 3.904761904761905e-05,
1808
+ "loss": 0.3842,
1809
+ "step": 2560
1810
+ },
1811
+ {
1812
+ "epoch": 1.8357142857142859,
1813
+ "grad_norm": 2.0849359035491943,
1814
+ "learning_rate": 3.880952380952381e-05,
1815
+ "loss": 0.2521,
1816
+ "step": 2570
1817
+ },
1818
+ {
1819
+ "epoch": 1.842857142857143,
1820
+ "grad_norm": 2.884455442428589,
1821
+ "learning_rate": 3.857142857142858e-05,
1822
+ "loss": 0.2784,
1823
+ "step": 2580
1824
+ },
1825
+ {
1826
+ "epoch": 1.85,
1827
+ "grad_norm": 2.0989837646484375,
1828
+ "learning_rate": 3.8333333333333334e-05,
1829
+ "loss": 0.2359,
1830
+ "step": 2590
1831
+ },
1832
+ {
1833
+ "epoch": 1.8571428571428572,
1834
+ "grad_norm": 3.074004650115967,
1835
+ "learning_rate": 3.809523809523809e-05,
1836
+ "loss": 0.2629,
1837
+ "step": 2600
1838
+ },
1839
+ {
1840
+ "epoch": 1.8642857142857143,
1841
+ "grad_norm": 3.4337799549102783,
1842
+ "learning_rate": 3.785714285714286e-05,
1843
+ "loss": 0.2449,
1844
+ "step": 2610
1845
+ },
1846
+ {
1847
+ "epoch": 1.8714285714285714,
1848
+ "grad_norm": 4.196152687072754,
1849
+ "learning_rate": 3.761904761904762e-05,
1850
+ "loss": 0.3082,
1851
+ "step": 2620
1852
+ },
1853
+ {
1854
+ "epoch": 1.8785714285714286,
1855
+ "grad_norm": 2.139364719390869,
1856
+ "learning_rate": 3.7380952380952386e-05,
1857
+ "loss": 0.2816,
1858
+ "step": 2630
1859
+ },
1860
+ {
1861
+ "epoch": 1.8857142857142857,
1862
+ "grad_norm": 6.65344762802124,
1863
+ "learning_rate": 3.7142857142857143e-05,
1864
+ "loss": 0.2729,
1865
+ "step": 2640
1866
+ },
1867
+ {
1868
+ "epoch": 1.8928571428571428,
1869
+ "grad_norm": 1.8867508172988892,
1870
+ "learning_rate": 3.690476190476191e-05,
1871
+ "loss": 0.2668,
1872
+ "step": 2650
1873
+ },
1874
+ {
1875
+ "epoch": 1.9,
1876
+ "grad_norm": 4.062267303466797,
1877
+ "learning_rate": 3.6666666666666666e-05,
1878
+ "loss": 0.2908,
1879
+ "step": 2660
1880
+ },
1881
+ {
1882
+ "epoch": 1.907142857142857,
1883
+ "grad_norm": 4.145746231079102,
1884
+ "learning_rate": 3.642857142857143e-05,
1885
+ "loss": 0.3168,
1886
+ "step": 2670
1887
+ },
1888
+ {
1889
+ "epoch": 1.9142857142857141,
1890
+ "grad_norm": 6.509150505065918,
1891
+ "learning_rate": 3.619047619047619e-05,
1892
+ "loss": 0.2744,
1893
+ "step": 2680
1894
+ },
1895
+ {
1896
+ "epoch": 1.9214285714285713,
1897
+ "grad_norm": 2.091609239578247,
1898
+ "learning_rate": 3.595238095238095e-05,
1899
+ "loss": 0.2945,
1900
+ "step": 2690
1901
+ },
1902
+ {
1903
+ "epoch": 1.9285714285714286,
1904
+ "grad_norm": 2.6082661151885986,
1905
+ "learning_rate": 3.571428571428572e-05,
1906
+ "loss": 0.2538,
1907
+ "step": 2700
1908
+ },
1909
+ {
1910
+ "epoch": 1.9357142857142857,
1911
+ "grad_norm": 3.4130637645721436,
1912
+ "learning_rate": 3.547619047619048e-05,
1913
+ "loss": 0.2726,
1914
+ "step": 2710
1915
+ },
1916
+ {
1917
+ "epoch": 1.9428571428571428,
1918
+ "grad_norm": 5.003239631652832,
1919
+ "learning_rate": 3.523809523809524e-05,
1920
+ "loss": 0.2697,
1921
+ "step": 2720
1922
+ },
1923
+ {
1924
+ "epoch": 1.95,
1925
+ "grad_norm": 2.7061688899993896,
1926
+ "learning_rate": 3.5e-05,
1927
+ "loss": 0.2756,
1928
+ "step": 2730
1929
+ },
1930
+ {
1931
+ "epoch": 1.9571428571428573,
1932
+ "grad_norm": 2.0089480876922607,
1933
+ "learning_rate": 3.476190476190476e-05,
1934
+ "loss": 0.2869,
1935
+ "step": 2740
1936
+ },
1937
+ {
1938
+ "epoch": 1.9642857142857144,
1939
+ "grad_norm": 2.2568588256835938,
1940
+ "learning_rate": 3.4523809523809526e-05,
1941
+ "loss": 0.2687,
1942
+ "step": 2750
1943
+ },
1944
+ {
1945
+ "epoch": 1.9714285714285715,
1946
+ "grad_norm": 2.0823915004730225,
1947
+ "learning_rate": 3.428571428571429e-05,
1948
+ "loss": 0.263,
1949
+ "step": 2760
1950
+ },
1951
+ {
1952
+ "epoch": 1.9785714285714286,
1953
+ "grad_norm": 1.738771915435791,
1954
+ "learning_rate": 3.404761904761905e-05,
1955
+ "loss": 0.2283,
1956
+ "step": 2770
1957
+ },
1958
+ {
1959
+ "epoch": 1.9857142857142858,
1960
+ "grad_norm": 4.0020623207092285,
1961
+ "learning_rate": 3.380952380952381e-05,
1962
+ "loss": 0.3286,
1963
+ "step": 2780
1964
+ },
1965
+ {
1966
+ "epoch": 1.9928571428571429,
1967
+ "grad_norm": 2.4708666801452637,
1968
+ "learning_rate": 3.357142857142857e-05,
1969
+ "loss": 0.279,
1970
+ "step": 2790
1971
+ },
1972
+ {
1973
+ "epoch": 2.0,
1974
+ "grad_norm": 1.950387716293335,
1975
+ "learning_rate": 3.3333333333333335e-05,
1976
+ "loss": 0.3078,
1977
+ "step": 2800
1978
+ },
1979
+ {
1980
+ "epoch": 2.0,
1981
+ "eval_loss": 0.25949418544769287,
1982
+ "eval_runtime": 68.2677,
1983
+ "eval_samples_per_second": 2.93,
1984
+ "eval_steps_per_second": 0.366,
1985
+ "step": 2800
1986
+ }
1987
+ ],
1988
+ "logging_steps": 10,
1989
+ "max_steps": 4200,
1990
+ "num_input_tokens_seen": 0,
1991
+ "num_train_epochs": 3,
1992
+ "save_steps": 500,
1993
+ "stateful_callbacks": {
1994
+ "TrainerControl": {
1995
+ "args": {
1996
+ "should_epoch_stop": false,
1997
+ "should_evaluate": false,
1998
+ "should_log": false,
1999
+ "should_save": true,
2000
+ "should_training_stop": false
2001
+ },
2002
+ "attributes": {}
2003
+ }
2004
+ },
2005
+ "total_flos": 1.6755883769856e+16,
2006
+ "train_batch_size": 2,
2007
+ "trial_name": null,
2008
+ "trial_params": null
2009
+ }
checkpoint-2800/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aef410e0e1d0110bb1523d95ae4f80f726beb1aea9164ada6672ae6ffeddc06b
3
+ size 5240
checkpoint-4200/README.md ADDED
@@ -0,0 +1,202 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ base_model: meta-llama/Llama-3.2-1B-Instruct
3
+ library_name: peft
4
+ ---
5
+
6
+ # Model Card for Model ID
7
+
8
+ <!-- Provide a quick summary of what the model is/does. -->
9
+
10
+
11
+
12
+ ## Model Details
13
+
14
+ ### Model Description
15
+
16
+ <!-- Provide a longer summary of what this model is. -->
17
+
18
+
19
+
20
+ - **Developed by:** [More Information Needed]
21
+ - **Funded by [optional]:** [More Information Needed]
22
+ - **Shared by [optional]:** [More Information Needed]
23
+ - **Model type:** [More Information Needed]
24
+ - **Language(s) (NLP):** [More Information Needed]
25
+ - **License:** [More Information Needed]
26
+ - **Finetuned from model [optional]:** [More Information Needed]
27
+
28
+ ### Model Sources [optional]
29
+
30
+ <!-- Provide the basic links for the model. -->
31
+
32
+ - **Repository:** [More Information Needed]
33
+ - **Paper [optional]:** [More Information Needed]
34
+ - **Demo [optional]:** [More Information Needed]
35
+
36
+ ## Uses
37
+
38
+ <!-- Address questions around how the model is intended to be used, including the foreseeable users of the model and those affected by the model. -->
39
+
40
+ ### Direct Use
41
+
42
+ <!-- This section is for the model use without fine-tuning or plugging into a larger ecosystem/app. -->
43
+
44
+ [More Information Needed]
45
+
46
+ ### Downstream Use [optional]
47
+
48
+ <!-- This section is for the model use when fine-tuned for a task, or when plugged into a larger ecosystem/app -->
49
+
50
+ [More Information Needed]
51
+
52
+ ### Out-of-Scope Use
53
+
54
+ <!-- This section addresses misuse, malicious use, and uses that the model will not work well for. -->
55
+
56
+ [More Information Needed]
57
+
58
+ ## Bias, Risks, and Limitations
59
+
60
+ <!-- This section is meant to convey both technical and sociotechnical limitations. -->
61
+
62
+ [More Information Needed]
63
+
64
+ ### Recommendations
65
+
66
+ <!-- This section is meant to convey recommendations with respect to the bias, risk, and technical limitations. -->
67
+
68
+ Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.
69
+
70
+ ## How to Get Started with the Model
71
+
72
+ Use the code below to get started with the model.
73
+
74
+ [More Information Needed]
75
+
76
+ ## Training Details
77
+
78
+ ### Training Data
79
+
80
+ <!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->
81
+
82
+ [More Information Needed]
83
+
84
+ ### Training Procedure
85
+
86
+ <!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->
87
+
88
+ #### Preprocessing [optional]
89
+
90
+ [More Information Needed]
91
+
92
+
93
+ #### Training Hyperparameters
94
+
95
+ - **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->
96
+
97
+ #### Speeds, Sizes, Times [optional]
98
+
99
+ <!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->
100
+
101
+ [More Information Needed]
102
+
103
+ ## Evaluation
104
+
105
+ <!-- This section describes the evaluation protocols and provides the results. -->
106
+
107
+ ### Testing Data, Factors & Metrics
108
+
109
+ #### Testing Data
110
+
111
+ <!-- This should link to a Dataset Card if possible. -->
112
+
113
+ [More Information Needed]
114
+
115
+ #### Factors
116
+
117
+ <!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->
118
+
119
+ [More Information Needed]
120
+
121
+ #### Metrics
122
+
123
+ <!-- These are the evaluation metrics being used, ideally with a description of why. -->
124
+
125
+ [More Information Needed]
126
+
127
+ ### Results
128
+
129
+ [More Information Needed]
130
+
131
+ #### Summary
132
+
133
+
134
+
135
+ ## Model Examination [optional]
136
+
137
+ <!-- Relevant interpretability work for the model goes here -->
138
+
139
+ [More Information Needed]
140
+
141
+ ## Environmental Impact
142
+
143
+ <!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->
144
+
145
+ Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).
146
+
147
+ - **Hardware Type:** [More Information Needed]
148
+ - **Hours used:** [More Information Needed]
149
+ - **Cloud Provider:** [More Information Needed]
150
+ - **Compute Region:** [More Information Needed]
151
+ - **Carbon Emitted:** [More Information Needed]
152
+
153
+ ## Technical Specifications [optional]
154
+
155
+ ### Model Architecture and Objective
156
+
157
+ [More Information Needed]
158
+
159
+ ### Compute Infrastructure
160
+
161
+ [More Information Needed]
162
+
163
+ #### Hardware
164
+
165
+ [More Information Needed]
166
+
167
+ #### Software
168
+
169
+ [More Information Needed]
170
+
171
+ ## Citation [optional]
172
+
173
+ <!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->
174
+
175
+ **BibTeX:**
176
+
177
+ [More Information Needed]
178
+
179
+ **APA:**
180
+
181
+ [More Information Needed]
182
+
183
+ ## Glossary [optional]
184
+
185
+ <!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->
186
+
187
+ [More Information Needed]
188
+
189
+ ## More Information [optional]
190
+
191
+ [More Information Needed]
192
+
193
+ ## Model Card Authors [optional]
194
+
195
+ [More Information Needed]
196
+
197
+ ## Model Card Contact
198
+
199
+ [More Information Needed]
200
+ ### Framework versions
201
+
202
+ - PEFT 0.13.2
checkpoint-4200/adapter_config.json ADDED
@@ -0,0 +1,29 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "alpha_pattern": {},
3
+ "auto_mapping": null,
4
+ "base_model_name_or_path": "meta-llama/Llama-3.2-1B-Instruct",
5
+ "bias": "none",
6
+ "fan_in_fan_out": false,
7
+ "inference_mode": true,
8
+ "init_lora_weights": true,
9
+ "layer_replication": null,
10
+ "layers_pattern": null,
11
+ "layers_to_transform": null,
12
+ "loftq_config": {},
13
+ "lora_alpha": 32,
14
+ "lora_dropout": 0.1,
15
+ "megatron_config": null,
16
+ "megatron_core": "megatron.core",
17
+ "modules_to_save": null,
18
+ "peft_type": "LORA",
19
+ "r": 8,
20
+ "rank_pattern": {},
21
+ "revision": null,
22
+ "target_modules": [
23
+ "q_proj",
24
+ "v_proj"
25
+ ],
26
+ "task_type": "CAUSAL_LM",
27
+ "use_dora": false,
28
+ "use_rslora": false
29
+ }
checkpoint-4200/adapter_model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ee723bec1693ba5b9d142a337187f000298d74d4c3f5e60a3af750f37abdd575
3
+ size 3416264
checkpoint-4200/optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:67458019e1bfcd2890756c9eb41e1b00a352d4398d2bb120a28691f55d12e532
3
+ size 6869818
checkpoint-4200/rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2944517e129c06a852c49794503048fba5f59a295db889beed8dbb938a2814b3
3
+ size 14244
checkpoint-4200/scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:74b3a38a5c66ed4f4680a26424471e34f122e55ecfba3a5fa5881090b6e78235
3
+ size 1064
checkpoint-4200/trainer_state.json ADDED
@@ -0,0 +1,2997 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_metric": null,
3
+ "best_model_checkpoint": null,
4
+ "epoch": 3.0,
5
+ "eval_steps": 500,
6
+ "global_step": 4200,
7
+ "is_hyper_param_search": false,
8
+ "is_local_process_zero": true,
9
+ "is_world_process_zero": true,
10
+ "log_history": [
11
+ {
12
+ "epoch": 0.007142857142857143,
13
+ "grad_norm": 2.519733190536499,
14
+ "learning_rate": 9.976190476190477e-05,
15
+ "loss": 1.2999,
16
+ "step": 10
17
+ },
18
+ {
19
+ "epoch": 0.014285714285714285,
20
+ "grad_norm": 2.4232161045074463,
21
+ "learning_rate": 9.952380952380953e-05,
22
+ "loss": 1.1228,
23
+ "step": 20
24
+ },
25
+ {
26
+ "epoch": 0.02142857142857143,
27
+ "grad_norm": 2.7126998901367188,
28
+ "learning_rate": 9.92857142857143e-05,
29
+ "loss": 0.9913,
30
+ "step": 30
31
+ },
32
+ {
33
+ "epoch": 0.02857142857142857,
34
+ "grad_norm": 7.3811235427856445,
35
+ "learning_rate": 9.904761904761905e-05,
36
+ "loss": 0.9306,
37
+ "step": 40
38
+ },
39
+ {
40
+ "epoch": 0.03571428571428571,
41
+ "grad_norm": 6.5697197914123535,
42
+ "learning_rate": 9.880952380952381e-05,
43
+ "loss": 0.7457,
44
+ "step": 50
45
+ },
46
+ {
47
+ "epoch": 0.04285714285714286,
48
+ "grad_norm": 1.2972687482833862,
49
+ "learning_rate": 9.857142857142858e-05,
50
+ "loss": 0.6408,
51
+ "step": 60
52
+ },
53
+ {
54
+ "epoch": 0.05,
55
+ "grad_norm": 3.600625991821289,
56
+ "learning_rate": 9.833333333333333e-05,
57
+ "loss": 0.6541,
58
+ "step": 70
59
+ },
60
+ {
61
+ "epoch": 0.05714285714285714,
62
+ "grad_norm": 1.5594621896743774,
63
+ "learning_rate": 9.80952380952381e-05,
64
+ "loss": 0.538,
65
+ "step": 80
66
+ },
67
+ {
68
+ "epoch": 0.06428571428571428,
69
+ "grad_norm": 4.595429420471191,
70
+ "learning_rate": 9.785714285714286e-05,
71
+ "loss": 0.556,
72
+ "step": 90
73
+ },
74
+ {
75
+ "epoch": 0.07142857142857142,
76
+ "grad_norm": 2.620643138885498,
77
+ "learning_rate": 9.761904761904762e-05,
78
+ "loss": 0.5521,
79
+ "step": 100
80
+ },
81
+ {
82
+ "epoch": 0.07857142857142857,
83
+ "grad_norm": 3.503641366958618,
84
+ "learning_rate": 9.738095238095239e-05,
85
+ "loss": 0.5912,
86
+ "step": 110
87
+ },
88
+ {
89
+ "epoch": 0.08571428571428572,
90
+ "grad_norm": 1.5058848857879639,
91
+ "learning_rate": 9.714285714285715e-05,
92
+ "loss": 0.5421,
93
+ "step": 120
94
+ },
95
+ {
96
+ "epoch": 0.09285714285714286,
97
+ "grad_norm": 3.7034990787506104,
98
+ "learning_rate": 9.69047619047619e-05,
99
+ "loss": 0.5568,
100
+ "step": 130
101
+ },
102
+ {
103
+ "epoch": 0.1,
104
+ "grad_norm": 3.714170455932617,
105
+ "learning_rate": 9.666666666666667e-05,
106
+ "loss": 0.5286,
107
+ "step": 140
108
+ },
109
+ {
110
+ "epoch": 0.10714285714285714,
111
+ "grad_norm": 2.069328546524048,
112
+ "learning_rate": 9.642857142857143e-05,
113
+ "loss": 0.4755,
114
+ "step": 150
115
+ },
116
+ {
117
+ "epoch": 0.11428571428571428,
118
+ "grad_norm": 1.2878583669662476,
119
+ "learning_rate": 9.61904761904762e-05,
120
+ "loss": 0.4817,
121
+ "step": 160
122
+ },
123
+ {
124
+ "epoch": 0.12142857142857143,
125
+ "grad_norm": 3.1945483684539795,
126
+ "learning_rate": 9.595238095238096e-05,
127
+ "loss": 0.5035,
128
+ "step": 170
129
+ },
130
+ {
131
+ "epoch": 0.12857142857142856,
132
+ "grad_norm": 1.9643166065216064,
133
+ "learning_rate": 9.571428571428573e-05,
134
+ "loss": 0.4932,
135
+ "step": 180
136
+ },
137
+ {
138
+ "epoch": 0.1357142857142857,
139
+ "grad_norm": 3.5002968311309814,
140
+ "learning_rate": 9.547619047619049e-05,
141
+ "loss": 0.5531,
142
+ "step": 190
143
+ },
144
+ {
145
+ "epoch": 0.14285714285714285,
146
+ "grad_norm": 4.7095794677734375,
147
+ "learning_rate": 9.523809523809524e-05,
148
+ "loss": 0.4601,
149
+ "step": 200
150
+ },
151
+ {
152
+ "epoch": 0.15,
153
+ "grad_norm": 1.4850050210952759,
154
+ "learning_rate": 9.5e-05,
155
+ "loss": 0.432,
156
+ "step": 210
157
+ },
158
+ {
159
+ "epoch": 0.15714285714285714,
160
+ "grad_norm": 2.860565423965454,
161
+ "learning_rate": 9.476190476190476e-05,
162
+ "loss": 0.4699,
163
+ "step": 220
164
+ },
165
+ {
166
+ "epoch": 0.16428571428571428,
167
+ "grad_norm": 1.8467603921890259,
168
+ "learning_rate": 9.452380952380952e-05,
169
+ "loss": 0.42,
170
+ "step": 230
171
+ },
172
+ {
173
+ "epoch": 0.17142857142857143,
174
+ "grad_norm": 1.6448352336883545,
175
+ "learning_rate": 9.428571428571429e-05,
176
+ "loss": 0.431,
177
+ "step": 240
178
+ },
179
+ {
180
+ "epoch": 0.17857142857142858,
181
+ "grad_norm": 1.8279732465744019,
182
+ "learning_rate": 9.404761904761905e-05,
183
+ "loss": 0.4413,
184
+ "step": 250
185
+ },
186
+ {
187
+ "epoch": 0.18571428571428572,
188
+ "grad_norm": 3.0112857818603516,
189
+ "learning_rate": 9.380952380952381e-05,
190
+ "loss": 0.4263,
191
+ "step": 260
192
+ },
193
+ {
194
+ "epoch": 0.19285714285714287,
195
+ "grad_norm": 1.7717293500900269,
196
+ "learning_rate": 9.357142857142858e-05,
197
+ "loss": 0.4829,
198
+ "step": 270
199
+ },
200
+ {
201
+ "epoch": 0.2,
202
+ "grad_norm": 4.515851974487305,
203
+ "learning_rate": 9.333333333333334e-05,
204
+ "loss": 0.4194,
205
+ "step": 280
206
+ },
207
+ {
208
+ "epoch": 0.20714285714285716,
209
+ "grad_norm": 3.998619556427002,
210
+ "learning_rate": 9.309523809523811e-05,
211
+ "loss": 0.4422,
212
+ "step": 290
213
+ },
214
+ {
215
+ "epoch": 0.21428571428571427,
216
+ "grad_norm": 3.5681912899017334,
217
+ "learning_rate": 9.285714285714286e-05,
218
+ "loss": 0.4677,
219
+ "step": 300
220
+ },
221
+ {
222
+ "epoch": 0.22142857142857142,
223
+ "grad_norm": 3.286348342895508,
224
+ "learning_rate": 9.261904761904762e-05,
225
+ "loss": 0.4152,
226
+ "step": 310
227
+ },
228
+ {
229
+ "epoch": 0.22857142857142856,
230
+ "grad_norm": 2.018721580505371,
231
+ "learning_rate": 9.238095238095239e-05,
232
+ "loss": 0.5065,
233
+ "step": 320
234
+ },
235
+ {
236
+ "epoch": 0.2357142857142857,
237
+ "grad_norm": 1.8489034175872803,
238
+ "learning_rate": 9.214285714285714e-05,
239
+ "loss": 0.392,
240
+ "step": 330
241
+ },
242
+ {
243
+ "epoch": 0.24285714285714285,
244
+ "grad_norm": 1.3958369493484497,
245
+ "learning_rate": 9.19047619047619e-05,
246
+ "loss": 0.3988,
247
+ "step": 340
248
+ },
249
+ {
250
+ "epoch": 0.25,
251
+ "grad_norm": 2.9772961139678955,
252
+ "learning_rate": 9.166666666666667e-05,
253
+ "loss": 0.3955,
254
+ "step": 350
255
+ },
256
+ {
257
+ "epoch": 0.2571428571428571,
258
+ "grad_norm": 3.2752914428710938,
259
+ "learning_rate": 9.142857142857143e-05,
260
+ "loss": 0.3998,
261
+ "step": 360
262
+ },
263
+ {
264
+ "epoch": 0.2642857142857143,
265
+ "grad_norm": 1.8811057806015015,
266
+ "learning_rate": 9.11904761904762e-05,
267
+ "loss": 0.4775,
268
+ "step": 370
269
+ },
270
+ {
271
+ "epoch": 0.2714285714285714,
272
+ "grad_norm": 2.3310513496398926,
273
+ "learning_rate": 9.095238095238096e-05,
274
+ "loss": 0.3703,
275
+ "step": 380
276
+ },
277
+ {
278
+ "epoch": 0.2785714285714286,
279
+ "grad_norm": 2.967597484588623,
280
+ "learning_rate": 9.071428571428571e-05,
281
+ "loss": 0.4177,
282
+ "step": 390
283
+ },
284
+ {
285
+ "epoch": 0.2857142857142857,
286
+ "grad_norm": 2.671673536300659,
287
+ "learning_rate": 9.047619047619048e-05,
288
+ "loss": 0.3823,
289
+ "step": 400
290
+ },
291
+ {
292
+ "epoch": 0.29285714285714287,
293
+ "grad_norm": 4.647867679595947,
294
+ "learning_rate": 9.023809523809524e-05,
295
+ "loss": 0.3888,
296
+ "step": 410
297
+ },
298
+ {
299
+ "epoch": 0.3,
300
+ "grad_norm": 3.4788432121276855,
301
+ "learning_rate": 9e-05,
302
+ "loss": 0.3856,
303
+ "step": 420
304
+ },
305
+ {
306
+ "epoch": 0.30714285714285716,
307
+ "grad_norm": 3.308420181274414,
308
+ "learning_rate": 8.976190476190477e-05,
309
+ "loss": 0.4143,
310
+ "step": 430
311
+ },
312
+ {
313
+ "epoch": 0.3142857142857143,
314
+ "grad_norm": 3.499880075454712,
315
+ "learning_rate": 8.952380952380953e-05,
316
+ "loss": 0.4857,
317
+ "step": 440
318
+ },
319
+ {
320
+ "epoch": 0.32142857142857145,
321
+ "grad_norm": 1.7907614707946777,
322
+ "learning_rate": 8.92857142857143e-05,
323
+ "loss": 0.356,
324
+ "step": 450
325
+ },
326
+ {
327
+ "epoch": 0.32857142857142857,
328
+ "grad_norm": 3.0111501216888428,
329
+ "learning_rate": 8.904761904761905e-05,
330
+ "loss": 0.358,
331
+ "step": 460
332
+ },
333
+ {
334
+ "epoch": 0.3357142857142857,
335
+ "grad_norm": 2.1389689445495605,
336
+ "learning_rate": 8.880952380952381e-05,
337
+ "loss": 0.3331,
338
+ "step": 470
339
+ },
340
+ {
341
+ "epoch": 0.34285714285714286,
342
+ "grad_norm": 1.8007326126098633,
343
+ "learning_rate": 8.857142857142857e-05,
344
+ "loss": 0.3148,
345
+ "step": 480
346
+ },
347
+ {
348
+ "epoch": 0.35,
349
+ "grad_norm": 3.603548288345337,
350
+ "learning_rate": 8.833333333333333e-05,
351
+ "loss": 0.4513,
352
+ "step": 490
353
+ },
354
+ {
355
+ "epoch": 0.35714285714285715,
356
+ "grad_norm": 2.270984411239624,
357
+ "learning_rate": 8.80952380952381e-05,
358
+ "loss": 0.3392,
359
+ "step": 500
360
+ },
361
+ {
362
+ "epoch": 0.36428571428571427,
363
+ "grad_norm": 1.5367858409881592,
364
+ "learning_rate": 8.785714285714286e-05,
365
+ "loss": 0.3665,
366
+ "step": 510
367
+ },
368
+ {
369
+ "epoch": 0.37142857142857144,
370
+ "grad_norm": 1.5397262573242188,
371
+ "learning_rate": 8.761904761904762e-05,
372
+ "loss": 0.3875,
373
+ "step": 520
374
+ },
375
+ {
376
+ "epoch": 0.37857142857142856,
377
+ "grad_norm": 3.6661577224731445,
378
+ "learning_rate": 8.738095238095239e-05,
379
+ "loss": 0.3617,
380
+ "step": 530
381
+ },
382
+ {
383
+ "epoch": 0.38571428571428573,
384
+ "grad_norm": 2.5946245193481445,
385
+ "learning_rate": 8.714285714285715e-05,
386
+ "loss": 0.3318,
387
+ "step": 540
388
+ },
389
+ {
390
+ "epoch": 0.39285714285714285,
391
+ "grad_norm": 2.508056640625,
392
+ "learning_rate": 8.690476190476192e-05,
393
+ "loss": 0.3681,
394
+ "step": 550
395
+ },
396
+ {
397
+ "epoch": 0.4,
398
+ "grad_norm": 5.406072616577148,
399
+ "learning_rate": 8.666666666666667e-05,
400
+ "loss": 0.3589,
401
+ "step": 560
402
+ },
403
+ {
404
+ "epoch": 0.40714285714285714,
405
+ "grad_norm": 2.0363543033599854,
406
+ "learning_rate": 8.642857142857143e-05,
407
+ "loss": 0.4328,
408
+ "step": 570
409
+ },
410
+ {
411
+ "epoch": 0.4142857142857143,
412
+ "grad_norm": 1.796236515045166,
413
+ "learning_rate": 8.61904761904762e-05,
414
+ "loss": 0.4005,
415
+ "step": 580
416
+ },
417
+ {
418
+ "epoch": 0.42142857142857143,
419
+ "grad_norm": 5.6172943115234375,
420
+ "learning_rate": 8.595238095238096e-05,
421
+ "loss": 0.3414,
422
+ "step": 590
423
+ },
424
+ {
425
+ "epoch": 0.42857142857142855,
426
+ "grad_norm": 2.953678846359253,
427
+ "learning_rate": 8.571428571428571e-05,
428
+ "loss": 0.3917,
429
+ "step": 600
430
+ },
431
+ {
432
+ "epoch": 0.4357142857142857,
433
+ "grad_norm": 1.8276687860488892,
434
+ "learning_rate": 8.547619047619048e-05,
435
+ "loss": 0.4057,
436
+ "step": 610
437
+ },
438
+ {
439
+ "epoch": 0.44285714285714284,
440
+ "grad_norm": 1.8668811321258545,
441
+ "learning_rate": 8.523809523809524e-05,
442
+ "loss": 0.4153,
443
+ "step": 620
444
+ },
445
+ {
446
+ "epoch": 0.45,
447
+ "grad_norm": 2.87809419631958,
448
+ "learning_rate": 8.5e-05,
449
+ "loss": 0.3911,
450
+ "step": 630
451
+ },
452
+ {
453
+ "epoch": 0.45714285714285713,
454
+ "grad_norm": 4.863674640655518,
455
+ "learning_rate": 8.476190476190477e-05,
456
+ "loss": 0.4458,
457
+ "step": 640
458
+ },
459
+ {
460
+ "epoch": 0.4642857142857143,
461
+ "grad_norm": 1.7150845527648926,
462
+ "learning_rate": 8.452380952380952e-05,
463
+ "loss": 0.3951,
464
+ "step": 650
465
+ },
466
+ {
467
+ "epoch": 0.4714285714285714,
468
+ "grad_norm": 2.0149145126342773,
469
+ "learning_rate": 8.428571428571429e-05,
470
+ "loss": 0.3238,
471
+ "step": 660
472
+ },
473
+ {
474
+ "epoch": 0.4785714285714286,
475
+ "grad_norm": 2.3146402835845947,
476
+ "learning_rate": 8.404761904761905e-05,
477
+ "loss": 0.3129,
478
+ "step": 670
479
+ },
480
+ {
481
+ "epoch": 0.4857142857142857,
482
+ "grad_norm": 1.8723537921905518,
483
+ "learning_rate": 8.380952380952382e-05,
484
+ "loss": 0.2969,
485
+ "step": 680
486
+ },
487
+ {
488
+ "epoch": 0.4928571428571429,
489
+ "grad_norm": 1.6206679344177246,
490
+ "learning_rate": 8.357142857142858e-05,
491
+ "loss": 0.3673,
492
+ "step": 690
493
+ },
494
+ {
495
+ "epoch": 0.5,
496
+ "grad_norm": 2.490722179412842,
497
+ "learning_rate": 8.333333333333334e-05,
498
+ "loss": 0.3493,
499
+ "step": 700
500
+ },
501
+ {
502
+ "epoch": 0.5071428571428571,
503
+ "grad_norm": 1.7868791818618774,
504
+ "learning_rate": 8.309523809523811e-05,
505
+ "loss": 0.423,
506
+ "step": 710
507
+ },
508
+ {
509
+ "epoch": 0.5142857142857142,
510
+ "grad_norm": 2.3152015209198,
511
+ "learning_rate": 8.285714285714287e-05,
512
+ "loss": 0.3448,
513
+ "step": 720
514
+ },
515
+ {
516
+ "epoch": 0.5214285714285715,
517
+ "grad_norm": 1.934523105621338,
518
+ "learning_rate": 8.261904761904762e-05,
519
+ "loss": 0.372,
520
+ "step": 730
521
+ },
522
+ {
523
+ "epoch": 0.5285714285714286,
524
+ "grad_norm": 2.421624183654785,
525
+ "learning_rate": 8.238095238095238e-05,
526
+ "loss": 0.3602,
527
+ "step": 740
528
+ },
529
+ {
530
+ "epoch": 0.5357142857142857,
531
+ "grad_norm": 2.0156285762786865,
532
+ "learning_rate": 8.214285714285714e-05,
533
+ "loss": 0.334,
534
+ "step": 750
535
+ },
536
+ {
537
+ "epoch": 0.5428571428571428,
538
+ "grad_norm": 1.6513855457305908,
539
+ "learning_rate": 8.19047619047619e-05,
540
+ "loss": 0.3553,
541
+ "step": 760
542
+ },
543
+ {
544
+ "epoch": 0.55,
545
+ "grad_norm": 1.950451374053955,
546
+ "learning_rate": 8.166666666666667e-05,
547
+ "loss": 0.3696,
548
+ "step": 770
549
+ },
550
+ {
551
+ "epoch": 0.5571428571428572,
552
+ "grad_norm": 2.4662423133850098,
553
+ "learning_rate": 8.142857142857143e-05,
554
+ "loss": 0.3232,
555
+ "step": 780
556
+ },
557
+ {
558
+ "epoch": 0.5642857142857143,
559
+ "grad_norm": 2.060654401779175,
560
+ "learning_rate": 8.11904761904762e-05,
561
+ "loss": 0.3294,
562
+ "step": 790
563
+ },
564
+ {
565
+ "epoch": 0.5714285714285714,
566
+ "grad_norm": 3.3581550121307373,
567
+ "learning_rate": 8.095238095238096e-05,
568
+ "loss": 0.3022,
569
+ "step": 800
570
+ },
571
+ {
572
+ "epoch": 0.5785714285714286,
573
+ "grad_norm": 1.6778148412704468,
574
+ "learning_rate": 8.071428571428573e-05,
575
+ "loss": 0.2808,
576
+ "step": 810
577
+ },
578
+ {
579
+ "epoch": 0.5857142857142857,
580
+ "grad_norm": 5.509958267211914,
581
+ "learning_rate": 8.047619047619048e-05,
582
+ "loss": 0.3584,
583
+ "step": 820
584
+ },
585
+ {
586
+ "epoch": 0.5928571428571429,
587
+ "grad_norm": 2.3136510848999023,
588
+ "learning_rate": 8.023809523809524e-05,
589
+ "loss": 0.2967,
590
+ "step": 830
591
+ },
592
+ {
593
+ "epoch": 0.6,
594
+ "grad_norm": 2.757749557495117,
595
+ "learning_rate": 8e-05,
596
+ "loss": 0.3091,
597
+ "step": 840
598
+ },
599
+ {
600
+ "epoch": 0.6071428571428571,
601
+ "grad_norm": 1.6453288793563843,
602
+ "learning_rate": 7.976190476190477e-05,
603
+ "loss": 0.2882,
604
+ "step": 850
605
+ },
606
+ {
607
+ "epoch": 0.6142857142857143,
608
+ "grad_norm": 3.599712371826172,
609
+ "learning_rate": 7.952380952380952e-05,
610
+ "loss": 0.35,
611
+ "step": 860
612
+ },
613
+ {
614
+ "epoch": 0.6214285714285714,
615
+ "grad_norm": 1.203050136566162,
616
+ "learning_rate": 7.928571428571429e-05,
617
+ "loss": 0.3147,
618
+ "step": 870
619
+ },
620
+ {
621
+ "epoch": 0.6285714285714286,
622
+ "grad_norm": 3.0582826137542725,
623
+ "learning_rate": 7.904761904761905e-05,
624
+ "loss": 0.369,
625
+ "step": 880
626
+ },
627
+ {
628
+ "epoch": 0.6357142857142857,
629
+ "grad_norm": 1.7133578062057495,
630
+ "learning_rate": 7.880952380952382e-05,
631
+ "loss": 0.4404,
632
+ "step": 890
633
+ },
634
+ {
635
+ "epoch": 0.6428571428571429,
636
+ "grad_norm": 2.1532819271087646,
637
+ "learning_rate": 7.857142857142858e-05,
638
+ "loss": 0.3249,
639
+ "step": 900
640
+ },
641
+ {
642
+ "epoch": 0.65,
643
+ "grad_norm": 3.529282808303833,
644
+ "learning_rate": 7.833333333333333e-05,
645
+ "loss": 0.3449,
646
+ "step": 910
647
+ },
648
+ {
649
+ "epoch": 0.6571428571428571,
650
+ "grad_norm": 3.412729501724243,
651
+ "learning_rate": 7.80952380952381e-05,
652
+ "loss": 0.3678,
653
+ "step": 920
654
+ },
655
+ {
656
+ "epoch": 0.6642857142857143,
657
+ "grad_norm": 1.8676691055297852,
658
+ "learning_rate": 7.785714285714286e-05,
659
+ "loss": 0.3522,
660
+ "step": 930
661
+ },
662
+ {
663
+ "epoch": 0.6714285714285714,
664
+ "grad_norm": 1.8748834133148193,
665
+ "learning_rate": 7.761904761904762e-05,
666
+ "loss": 0.2561,
667
+ "step": 940
668
+ },
669
+ {
670
+ "epoch": 0.6785714285714286,
671
+ "grad_norm": 3.1622543334960938,
672
+ "learning_rate": 7.738095238095239e-05,
673
+ "loss": 0.3245,
674
+ "step": 950
675
+ },
676
+ {
677
+ "epoch": 0.6857142857142857,
678
+ "grad_norm": 2.17108416557312,
679
+ "learning_rate": 7.714285714285715e-05,
680
+ "loss": 0.335,
681
+ "step": 960
682
+ },
683
+ {
684
+ "epoch": 0.6928571428571428,
685
+ "grad_norm": 3.686673402786255,
686
+ "learning_rate": 7.690476190476192e-05,
687
+ "loss": 0.319,
688
+ "step": 970
689
+ },
690
+ {
691
+ "epoch": 0.7,
692
+ "grad_norm": 7.108774662017822,
693
+ "learning_rate": 7.666666666666667e-05,
694
+ "loss": 0.3337,
695
+ "step": 980
696
+ },
697
+ {
698
+ "epoch": 0.7071428571428572,
699
+ "grad_norm": 2.532336711883545,
700
+ "learning_rate": 7.642857142857143e-05,
701
+ "loss": 0.377,
702
+ "step": 990
703
+ },
704
+ {
705
+ "epoch": 0.7142857142857143,
706
+ "grad_norm": 1.8768038749694824,
707
+ "learning_rate": 7.619047619047618e-05,
708
+ "loss": 0.3745,
709
+ "step": 1000
710
+ },
711
+ {
712
+ "epoch": 0.7214285714285714,
713
+ "grad_norm": 1.8926101922988892,
714
+ "learning_rate": 7.595238095238095e-05,
715
+ "loss": 0.3503,
716
+ "step": 1010
717
+ },
718
+ {
719
+ "epoch": 0.7285714285714285,
720
+ "grad_norm": 2.2763147354125977,
721
+ "learning_rate": 7.571428571428571e-05,
722
+ "loss": 0.307,
723
+ "step": 1020
724
+ },
725
+ {
726
+ "epoch": 0.7357142857142858,
727
+ "grad_norm": 3.663257360458374,
728
+ "learning_rate": 7.547619047619048e-05,
729
+ "loss": 0.2876,
730
+ "step": 1030
731
+ },
732
+ {
733
+ "epoch": 0.7428571428571429,
734
+ "grad_norm": 1.844648003578186,
735
+ "learning_rate": 7.523809523809524e-05,
736
+ "loss": 0.3512,
737
+ "step": 1040
738
+ },
739
+ {
740
+ "epoch": 0.75,
741
+ "grad_norm": 2.5800540447235107,
742
+ "learning_rate": 7.500000000000001e-05,
743
+ "loss": 0.301,
744
+ "step": 1050
745
+ },
746
+ {
747
+ "epoch": 0.7571428571428571,
748
+ "grad_norm": 2.4674232006073,
749
+ "learning_rate": 7.476190476190477e-05,
750
+ "loss": 0.3635,
751
+ "step": 1060
752
+ },
753
+ {
754
+ "epoch": 0.7642857142857142,
755
+ "grad_norm": 3.5119080543518066,
756
+ "learning_rate": 7.452380952380952e-05,
757
+ "loss": 0.3688,
758
+ "step": 1070
759
+ },
760
+ {
761
+ "epoch": 0.7714285714285715,
762
+ "grad_norm": 2.5902345180511475,
763
+ "learning_rate": 7.428571428571429e-05,
764
+ "loss": 0.3136,
765
+ "step": 1080
766
+ },
767
+ {
768
+ "epoch": 0.7785714285714286,
769
+ "grad_norm": 1.734082818031311,
770
+ "learning_rate": 7.404761904761905e-05,
771
+ "loss": 0.3366,
772
+ "step": 1090
773
+ },
774
+ {
775
+ "epoch": 0.7857142857142857,
776
+ "grad_norm": 2.0863406658172607,
777
+ "learning_rate": 7.380952380952382e-05,
778
+ "loss": 0.331,
779
+ "step": 1100
780
+ },
781
+ {
782
+ "epoch": 0.7928571428571428,
783
+ "grad_norm": 2.9227652549743652,
784
+ "learning_rate": 7.357142857142858e-05,
785
+ "loss": 0.3131,
786
+ "step": 1110
787
+ },
788
+ {
789
+ "epoch": 0.8,
790
+ "grad_norm": 5.268605709075928,
791
+ "learning_rate": 7.333333333333333e-05,
792
+ "loss": 0.3212,
793
+ "step": 1120
794
+ },
795
+ {
796
+ "epoch": 0.8071428571428572,
797
+ "grad_norm": 2.633431911468506,
798
+ "learning_rate": 7.30952380952381e-05,
799
+ "loss": 0.2707,
800
+ "step": 1130
801
+ },
802
+ {
803
+ "epoch": 0.8142857142857143,
804
+ "grad_norm": 1.652354121208191,
805
+ "learning_rate": 7.285714285714286e-05,
806
+ "loss": 0.3037,
807
+ "step": 1140
808
+ },
809
+ {
810
+ "epoch": 0.8214285714285714,
811
+ "grad_norm": 3.1987786293029785,
812
+ "learning_rate": 7.261904761904762e-05,
813
+ "loss": 0.3603,
814
+ "step": 1150
815
+ },
816
+ {
817
+ "epoch": 0.8285714285714286,
818
+ "grad_norm": 1.7624555826187134,
819
+ "learning_rate": 7.238095238095238e-05,
820
+ "loss": 0.308,
821
+ "step": 1160
822
+ },
823
+ {
824
+ "epoch": 0.8357142857142857,
825
+ "grad_norm": 2.6379318237304688,
826
+ "learning_rate": 7.214285714285714e-05,
827
+ "loss": 0.3068,
828
+ "step": 1170
829
+ },
830
+ {
831
+ "epoch": 0.8428571428571429,
832
+ "grad_norm": 2.6688101291656494,
833
+ "learning_rate": 7.19047619047619e-05,
834
+ "loss": 0.3134,
835
+ "step": 1180
836
+ },
837
+ {
838
+ "epoch": 0.85,
839
+ "grad_norm": 3.698643207550049,
840
+ "learning_rate": 7.166666666666667e-05,
841
+ "loss": 0.2973,
842
+ "step": 1190
843
+ },
844
+ {
845
+ "epoch": 0.8571428571428571,
846
+ "grad_norm": 2.9004178047180176,
847
+ "learning_rate": 7.142857142857143e-05,
848
+ "loss": 0.3116,
849
+ "step": 1200
850
+ },
851
+ {
852
+ "epoch": 0.8642857142857143,
853
+ "grad_norm": 2.3414883613586426,
854
+ "learning_rate": 7.11904761904762e-05,
855
+ "loss": 0.334,
856
+ "step": 1210
857
+ },
858
+ {
859
+ "epoch": 0.8714285714285714,
860
+ "grad_norm": 3.6203651428222656,
861
+ "learning_rate": 7.095238095238096e-05,
862
+ "loss": 0.3241,
863
+ "step": 1220
864
+ },
865
+ {
866
+ "epoch": 0.8785714285714286,
867
+ "grad_norm": 2.7472829818725586,
868
+ "learning_rate": 7.071428571428573e-05,
869
+ "loss": 0.2727,
870
+ "step": 1230
871
+ },
872
+ {
873
+ "epoch": 0.8857142857142857,
874
+ "grad_norm": 3.2332115173339844,
875
+ "learning_rate": 7.047619047619048e-05,
876
+ "loss": 0.3329,
877
+ "step": 1240
878
+ },
879
+ {
880
+ "epoch": 0.8928571428571429,
881
+ "grad_norm": 1.968595027923584,
882
+ "learning_rate": 7.023809523809524e-05,
883
+ "loss": 0.3478,
884
+ "step": 1250
885
+ },
886
+ {
887
+ "epoch": 0.9,
888
+ "grad_norm": 2.3712966442108154,
889
+ "learning_rate": 7e-05,
890
+ "loss": 0.3964,
891
+ "step": 1260
892
+ },
893
+ {
894
+ "epoch": 0.9071428571428571,
895
+ "grad_norm": 1.7704048156738281,
896
+ "learning_rate": 6.976190476190476e-05,
897
+ "loss": 0.3983,
898
+ "step": 1270
899
+ },
900
+ {
901
+ "epoch": 0.9142857142857143,
902
+ "grad_norm": 1.8879201412200928,
903
+ "learning_rate": 6.952380952380952e-05,
904
+ "loss": 0.3296,
905
+ "step": 1280
906
+ },
907
+ {
908
+ "epoch": 0.9214285714285714,
909
+ "grad_norm": 4.0522780418396,
910
+ "learning_rate": 6.928571428571429e-05,
911
+ "loss": 0.3722,
912
+ "step": 1290
913
+ },
914
+ {
915
+ "epoch": 0.9285714285714286,
916
+ "grad_norm": 1.7849159240722656,
917
+ "learning_rate": 6.904761904761905e-05,
918
+ "loss": 0.3022,
919
+ "step": 1300
920
+ },
921
+ {
922
+ "epoch": 0.9357142857142857,
923
+ "grad_norm": 3.376678705215454,
924
+ "learning_rate": 6.880952380952382e-05,
925
+ "loss": 0.3429,
926
+ "step": 1310
927
+ },
928
+ {
929
+ "epoch": 0.9428571428571428,
930
+ "grad_norm": 4.100617408752441,
931
+ "learning_rate": 6.857142857142858e-05,
932
+ "loss": 0.2793,
933
+ "step": 1320
934
+ },
935
+ {
936
+ "epoch": 0.95,
937
+ "grad_norm": 3.8000288009643555,
938
+ "learning_rate": 6.833333333333333e-05,
939
+ "loss": 0.3114,
940
+ "step": 1330
941
+ },
942
+ {
943
+ "epoch": 0.9571428571428572,
944
+ "grad_norm": 4.846517562866211,
945
+ "learning_rate": 6.80952380952381e-05,
946
+ "loss": 0.3377,
947
+ "step": 1340
948
+ },
949
+ {
950
+ "epoch": 0.9642857142857143,
951
+ "grad_norm": 2.0567409992218018,
952
+ "learning_rate": 6.785714285714286e-05,
953
+ "loss": 0.3198,
954
+ "step": 1350
955
+ },
956
+ {
957
+ "epoch": 0.9714285714285714,
958
+ "grad_norm": 2.3197948932647705,
959
+ "learning_rate": 6.761904761904763e-05,
960
+ "loss": 0.2817,
961
+ "step": 1360
962
+ },
963
+ {
964
+ "epoch": 0.9785714285714285,
965
+ "grad_norm": 2.7312827110290527,
966
+ "learning_rate": 6.738095238095239e-05,
967
+ "loss": 0.3378,
968
+ "step": 1370
969
+ },
970
+ {
971
+ "epoch": 0.9857142857142858,
972
+ "grad_norm": 3.115516185760498,
973
+ "learning_rate": 6.714285714285714e-05,
974
+ "loss": 0.3638,
975
+ "step": 1380
976
+ },
977
+ {
978
+ "epoch": 0.9928571428571429,
979
+ "grad_norm": 1.0742294788360596,
980
+ "learning_rate": 6.69047619047619e-05,
981
+ "loss": 0.3176,
982
+ "step": 1390
983
+ },
984
+ {
985
+ "epoch": 1.0,
986
+ "grad_norm": 2.3872761726379395,
987
+ "learning_rate": 6.666666666666667e-05,
988
+ "loss": 0.2921,
989
+ "step": 1400
990
+ },
991
+ {
992
+ "epoch": 1.0,
993
+ "eval_loss": 0.2929375469684601,
994
+ "eval_runtime": 68.5646,
995
+ "eval_samples_per_second": 2.917,
996
+ "eval_steps_per_second": 0.365,
997
+ "step": 1400
998
+ },
999
+ {
1000
+ "epoch": 1.0071428571428571,
1001
+ "grad_norm": 2.1662676334381104,
1002
+ "learning_rate": 6.642857142857143e-05,
1003
+ "loss": 0.2885,
1004
+ "step": 1410
1005
+ },
1006
+ {
1007
+ "epoch": 1.0142857142857142,
1008
+ "grad_norm": 4.116220474243164,
1009
+ "learning_rate": 6.619047619047619e-05,
1010
+ "loss": 0.2951,
1011
+ "step": 1420
1012
+ },
1013
+ {
1014
+ "epoch": 1.0214285714285714,
1015
+ "grad_norm": 4.550313472747803,
1016
+ "learning_rate": 6.595238095238095e-05,
1017
+ "loss": 0.3078,
1018
+ "step": 1430
1019
+ },
1020
+ {
1021
+ "epoch": 1.0285714285714285,
1022
+ "grad_norm": 1.7829937934875488,
1023
+ "learning_rate": 6.571428571428571e-05,
1024
+ "loss": 0.3309,
1025
+ "step": 1440
1026
+ },
1027
+ {
1028
+ "epoch": 1.0357142857142858,
1029
+ "grad_norm": 2.684500217437744,
1030
+ "learning_rate": 6.547619047619048e-05,
1031
+ "loss": 0.2994,
1032
+ "step": 1450
1033
+ },
1034
+ {
1035
+ "epoch": 1.042857142857143,
1036
+ "grad_norm": 1.3234150409698486,
1037
+ "learning_rate": 6.523809523809524e-05,
1038
+ "loss": 0.3782,
1039
+ "step": 1460
1040
+ },
1041
+ {
1042
+ "epoch": 1.05,
1043
+ "grad_norm": 2.6253223419189453,
1044
+ "learning_rate": 6.500000000000001e-05,
1045
+ "loss": 0.3094,
1046
+ "step": 1470
1047
+ },
1048
+ {
1049
+ "epoch": 1.0571428571428572,
1050
+ "grad_norm": 3.272584915161133,
1051
+ "learning_rate": 6.476190476190477e-05,
1052
+ "loss": 0.3406,
1053
+ "step": 1480
1054
+ },
1055
+ {
1056
+ "epoch": 1.0642857142857143,
1057
+ "grad_norm": 3.765272378921509,
1058
+ "learning_rate": 6.452380952380954e-05,
1059
+ "loss": 0.3525,
1060
+ "step": 1490
1061
+ },
1062
+ {
1063
+ "epoch": 1.0714285714285714,
1064
+ "grad_norm": 4.5847578048706055,
1065
+ "learning_rate": 6.428571428571429e-05,
1066
+ "loss": 0.3296,
1067
+ "step": 1500
1068
+ },
1069
+ {
1070
+ "epoch": 1.0785714285714285,
1071
+ "grad_norm": 2.448350191116333,
1072
+ "learning_rate": 6.404761904761904e-05,
1073
+ "loss": 0.2704,
1074
+ "step": 1510
1075
+ },
1076
+ {
1077
+ "epoch": 1.0857142857142856,
1078
+ "grad_norm": 2.5641868114471436,
1079
+ "learning_rate": 6.38095238095238e-05,
1080
+ "loss": 0.277,
1081
+ "step": 1520
1082
+ },
1083
+ {
1084
+ "epoch": 1.092857142857143,
1085
+ "grad_norm": 1.8631259202957153,
1086
+ "learning_rate": 6.357142857142857e-05,
1087
+ "loss": 0.2915,
1088
+ "step": 1530
1089
+ },
1090
+ {
1091
+ "epoch": 1.1,
1092
+ "grad_norm": 1.7823829650878906,
1093
+ "learning_rate": 6.333333333333333e-05,
1094
+ "loss": 0.279,
1095
+ "step": 1540
1096
+ },
1097
+ {
1098
+ "epoch": 1.1071428571428572,
1099
+ "grad_norm": 3.2730162143707275,
1100
+ "learning_rate": 6.30952380952381e-05,
1101
+ "loss": 0.3179,
1102
+ "step": 1550
1103
+ },
1104
+ {
1105
+ "epoch": 1.1142857142857143,
1106
+ "grad_norm": 2.495410203933716,
1107
+ "learning_rate": 6.285714285714286e-05,
1108
+ "loss": 0.326,
1109
+ "step": 1560
1110
+ },
1111
+ {
1112
+ "epoch": 1.1214285714285714,
1113
+ "grad_norm": 1.844226598739624,
1114
+ "learning_rate": 6.261904761904763e-05,
1115
+ "loss": 0.2931,
1116
+ "step": 1570
1117
+ },
1118
+ {
1119
+ "epoch": 1.1285714285714286,
1120
+ "grad_norm": 3.5361154079437256,
1121
+ "learning_rate": 6.238095238095239e-05,
1122
+ "loss": 0.2932,
1123
+ "step": 1580
1124
+ },
1125
+ {
1126
+ "epoch": 1.1357142857142857,
1127
+ "grad_norm": 1.9047455787658691,
1128
+ "learning_rate": 6.214285714285714e-05,
1129
+ "loss": 0.2779,
1130
+ "step": 1590
1131
+ },
1132
+ {
1133
+ "epoch": 1.1428571428571428,
1134
+ "grad_norm": 6.630354881286621,
1135
+ "learning_rate": 6.19047619047619e-05,
1136
+ "loss": 0.3563,
1137
+ "step": 1600
1138
+ },
1139
+ {
1140
+ "epoch": 1.15,
1141
+ "grad_norm": 4.267199993133545,
1142
+ "learning_rate": 6.166666666666667e-05,
1143
+ "loss": 0.3146,
1144
+ "step": 1610
1145
+ },
1146
+ {
1147
+ "epoch": 1.157142857142857,
1148
+ "grad_norm": 4.094973087310791,
1149
+ "learning_rate": 6.142857142857143e-05,
1150
+ "loss": 0.3096,
1151
+ "step": 1620
1152
+ },
1153
+ {
1154
+ "epoch": 1.1642857142857144,
1155
+ "grad_norm": 2.1625826358795166,
1156
+ "learning_rate": 6.11904761904762e-05,
1157
+ "loss": 0.2893,
1158
+ "step": 1630
1159
+ },
1160
+ {
1161
+ "epoch": 1.1714285714285715,
1162
+ "grad_norm": 4.5134758949279785,
1163
+ "learning_rate": 6.0952380952380964e-05,
1164
+ "loss": 0.3037,
1165
+ "step": 1640
1166
+ },
1167
+ {
1168
+ "epoch": 1.1785714285714286,
1169
+ "grad_norm": 1.987043023109436,
1170
+ "learning_rate": 6.0714285714285715e-05,
1171
+ "loss": 0.3427,
1172
+ "step": 1650
1173
+ },
1174
+ {
1175
+ "epoch": 1.1857142857142857,
1176
+ "grad_norm": 1.7972902059555054,
1177
+ "learning_rate": 6.047619047619047e-05,
1178
+ "loss": 0.3012,
1179
+ "step": 1660
1180
+ },
1181
+ {
1182
+ "epoch": 1.1928571428571428,
1183
+ "grad_norm": 2.2862155437469482,
1184
+ "learning_rate": 6.023809523809524e-05,
1185
+ "loss": 0.2877,
1186
+ "step": 1670
1187
+ },
1188
+ {
1189
+ "epoch": 1.2,
1190
+ "grad_norm": 2.5027403831481934,
1191
+ "learning_rate": 6e-05,
1192
+ "loss": 0.3172,
1193
+ "step": 1680
1194
+ },
1195
+ {
1196
+ "epoch": 1.207142857142857,
1197
+ "grad_norm": 2.698021173477173,
1198
+ "learning_rate": 5.9761904761904766e-05,
1199
+ "loss": 0.2689,
1200
+ "step": 1690
1201
+ },
1202
+ {
1203
+ "epoch": 1.2142857142857142,
1204
+ "grad_norm": 1.9757822751998901,
1205
+ "learning_rate": 5.9523809523809524e-05,
1206
+ "loss": 0.2845,
1207
+ "step": 1700
1208
+ },
1209
+ {
1210
+ "epoch": 1.2214285714285715,
1211
+ "grad_norm": 1.7587430477142334,
1212
+ "learning_rate": 5.928571428571429e-05,
1213
+ "loss": 0.3264,
1214
+ "step": 1710
1215
+ },
1216
+ {
1217
+ "epoch": 1.2285714285714286,
1218
+ "grad_norm": 2.466726541519165,
1219
+ "learning_rate": 5.904761904761905e-05,
1220
+ "loss": 0.2311,
1221
+ "step": 1720
1222
+ },
1223
+ {
1224
+ "epoch": 1.2357142857142858,
1225
+ "grad_norm": 2.041752815246582,
1226
+ "learning_rate": 5.880952380952382e-05,
1227
+ "loss": 0.2829,
1228
+ "step": 1730
1229
+ },
1230
+ {
1231
+ "epoch": 1.2428571428571429,
1232
+ "grad_norm": 4.565363883972168,
1233
+ "learning_rate": 5.8571428571428575e-05,
1234
+ "loss": 0.2356,
1235
+ "step": 1740
1236
+ },
1237
+ {
1238
+ "epoch": 1.25,
1239
+ "grad_norm": 3.1162221431732178,
1240
+ "learning_rate": 5.833333333333334e-05,
1241
+ "loss": 0.2719,
1242
+ "step": 1750
1243
+ },
1244
+ {
1245
+ "epoch": 1.2571428571428571,
1246
+ "grad_norm": 4.6578779220581055,
1247
+ "learning_rate": 5.8095238095238104e-05,
1248
+ "loss": 0.2477,
1249
+ "step": 1760
1250
+ },
1251
+ {
1252
+ "epoch": 1.2642857142857142,
1253
+ "grad_norm": 2.782367706298828,
1254
+ "learning_rate": 5.785714285714287e-05,
1255
+ "loss": 0.3718,
1256
+ "step": 1770
1257
+ },
1258
+ {
1259
+ "epoch": 1.2714285714285714,
1260
+ "grad_norm": 3.4065630435943604,
1261
+ "learning_rate": 5.761904761904762e-05,
1262
+ "loss": 0.3211,
1263
+ "step": 1780
1264
+ },
1265
+ {
1266
+ "epoch": 1.2785714285714285,
1267
+ "grad_norm": 2.6382477283477783,
1268
+ "learning_rate": 5.738095238095238e-05,
1269
+ "loss": 0.3476,
1270
+ "step": 1790
1271
+ },
1272
+ {
1273
+ "epoch": 1.2857142857142856,
1274
+ "grad_norm": 1.6404658555984497,
1275
+ "learning_rate": 5.714285714285714e-05,
1276
+ "loss": 0.3465,
1277
+ "step": 1800
1278
+ },
1279
+ {
1280
+ "epoch": 1.292857142857143,
1281
+ "grad_norm": 2.675689935684204,
1282
+ "learning_rate": 5.6904761904761906e-05,
1283
+ "loss": 0.3027,
1284
+ "step": 1810
1285
+ },
1286
+ {
1287
+ "epoch": 1.3,
1288
+ "grad_norm": 3.062359571456909,
1289
+ "learning_rate": 5.666666666666667e-05,
1290
+ "loss": 0.3043,
1291
+ "step": 1820
1292
+ },
1293
+ {
1294
+ "epoch": 1.3071428571428572,
1295
+ "grad_norm": 3.5478060245513916,
1296
+ "learning_rate": 5.642857142857143e-05,
1297
+ "loss": 0.3022,
1298
+ "step": 1830
1299
+ },
1300
+ {
1301
+ "epoch": 1.3142857142857143,
1302
+ "grad_norm": 3.2227487564086914,
1303
+ "learning_rate": 5.619047619047619e-05,
1304
+ "loss": 0.2776,
1305
+ "step": 1840
1306
+ },
1307
+ {
1308
+ "epoch": 1.3214285714285714,
1309
+ "grad_norm": 1.7066378593444824,
1310
+ "learning_rate": 5.595238095238096e-05,
1311
+ "loss": 0.3345,
1312
+ "step": 1850
1313
+ },
1314
+ {
1315
+ "epoch": 1.3285714285714285,
1316
+ "grad_norm": 3.7412548065185547,
1317
+ "learning_rate": 5.571428571428572e-05,
1318
+ "loss": 0.2886,
1319
+ "step": 1860
1320
+ },
1321
+ {
1322
+ "epoch": 1.3357142857142856,
1323
+ "grad_norm": 6.295001983642578,
1324
+ "learning_rate": 5.547619047619048e-05,
1325
+ "loss": 0.2926,
1326
+ "step": 1870
1327
+ },
1328
+ {
1329
+ "epoch": 1.342857142857143,
1330
+ "grad_norm": 2.9826364517211914,
1331
+ "learning_rate": 5.5238095238095244e-05,
1332
+ "loss": 0.2908,
1333
+ "step": 1880
1334
+ },
1335
+ {
1336
+ "epoch": 1.35,
1337
+ "grad_norm": 2.300792932510376,
1338
+ "learning_rate": 5.500000000000001e-05,
1339
+ "loss": 0.3346,
1340
+ "step": 1890
1341
+ },
1342
+ {
1343
+ "epoch": 1.3571428571428572,
1344
+ "grad_norm": 3.02064847946167,
1345
+ "learning_rate": 5.4761904761904766e-05,
1346
+ "loss": 0.298,
1347
+ "step": 1900
1348
+ },
1349
+ {
1350
+ "epoch": 1.3642857142857143,
1351
+ "grad_norm": 2.3579671382904053,
1352
+ "learning_rate": 5.4523809523809524e-05,
1353
+ "loss": 0.3148,
1354
+ "step": 1910
1355
+ },
1356
+ {
1357
+ "epoch": 1.3714285714285714,
1358
+ "grad_norm": 2.4837729930877686,
1359
+ "learning_rate": 5.428571428571428e-05,
1360
+ "loss": 0.2942,
1361
+ "step": 1920
1362
+ },
1363
+ {
1364
+ "epoch": 1.3785714285714286,
1365
+ "grad_norm": 1.5465370416641235,
1366
+ "learning_rate": 5.4047619047619046e-05,
1367
+ "loss": 0.3141,
1368
+ "step": 1930
1369
+ },
1370
+ {
1371
+ "epoch": 1.3857142857142857,
1372
+ "grad_norm": 3.10286808013916,
1373
+ "learning_rate": 5.380952380952381e-05,
1374
+ "loss": 0.2519,
1375
+ "step": 1940
1376
+ },
1377
+ {
1378
+ "epoch": 1.3928571428571428,
1379
+ "grad_norm": 2.024423837661743,
1380
+ "learning_rate": 5.3571428571428575e-05,
1381
+ "loss": 0.2399,
1382
+ "step": 1950
1383
+ },
1384
+ {
1385
+ "epoch": 1.4,
1386
+ "grad_norm": 3.297781467437744,
1387
+ "learning_rate": 5.333333333333333e-05,
1388
+ "loss": 0.3115,
1389
+ "step": 1960
1390
+ },
1391
+ {
1392
+ "epoch": 1.407142857142857,
1393
+ "grad_norm": 12.920825004577637,
1394
+ "learning_rate": 5.30952380952381e-05,
1395
+ "loss": 0.3199,
1396
+ "step": 1970
1397
+ },
1398
+ {
1399
+ "epoch": 1.4142857142857144,
1400
+ "grad_norm": 1.6162301301956177,
1401
+ "learning_rate": 5.285714285714286e-05,
1402
+ "loss": 0.228,
1403
+ "step": 1980
1404
+ },
1405
+ {
1406
+ "epoch": 1.4214285714285715,
1407
+ "grad_norm": 3.950810194015503,
1408
+ "learning_rate": 5.261904761904763e-05,
1409
+ "loss": 0.3363,
1410
+ "step": 1990
1411
+ },
1412
+ {
1413
+ "epoch": 1.4285714285714286,
1414
+ "grad_norm": 1.5367783308029175,
1415
+ "learning_rate": 5.2380952380952384e-05,
1416
+ "loss": 0.2328,
1417
+ "step": 2000
1418
+ },
1419
+ {
1420
+ "epoch": 1.4357142857142857,
1421
+ "grad_norm": 1.4729127883911133,
1422
+ "learning_rate": 5.214285714285715e-05,
1423
+ "loss": 0.2892,
1424
+ "step": 2010
1425
+ },
1426
+ {
1427
+ "epoch": 1.4428571428571428,
1428
+ "grad_norm": 3.7027993202209473,
1429
+ "learning_rate": 5.1904761904761913e-05,
1430
+ "loss": 0.2598,
1431
+ "step": 2020
1432
+ },
1433
+ {
1434
+ "epoch": 1.45,
1435
+ "grad_norm": 4.046818256378174,
1436
+ "learning_rate": 5.166666666666667e-05,
1437
+ "loss": 0.3184,
1438
+ "step": 2030
1439
+ },
1440
+ {
1441
+ "epoch": 1.457142857142857,
1442
+ "grad_norm": 5.93725061416626,
1443
+ "learning_rate": 5.142857142857143e-05,
1444
+ "loss": 0.2777,
1445
+ "step": 2040
1446
+ },
1447
+ {
1448
+ "epoch": 1.4642857142857144,
1449
+ "grad_norm": 2.02740740776062,
1450
+ "learning_rate": 5.119047619047619e-05,
1451
+ "loss": 0.2632,
1452
+ "step": 2050
1453
+ },
1454
+ {
1455
+ "epoch": 1.4714285714285715,
1456
+ "grad_norm": 1.166384220123291,
1457
+ "learning_rate": 5.095238095238095e-05,
1458
+ "loss": 0.2457,
1459
+ "step": 2060
1460
+ },
1461
+ {
1462
+ "epoch": 1.4785714285714286,
1463
+ "grad_norm": 2.251250982284546,
1464
+ "learning_rate": 5.0714285714285716e-05,
1465
+ "loss": 0.2996,
1466
+ "step": 2070
1467
+ },
1468
+ {
1469
+ "epoch": 1.4857142857142858,
1470
+ "grad_norm": 1.698370337486267,
1471
+ "learning_rate": 5.047619047619048e-05,
1472
+ "loss": 0.2418,
1473
+ "step": 2080
1474
+ },
1475
+ {
1476
+ "epoch": 1.4928571428571429,
1477
+ "grad_norm": 1.665677785873413,
1478
+ "learning_rate": 5.023809523809524e-05,
1479
+ "loss": 0.2718,
1480
+ "step": 2090
1481
+ },
1482
+ {
1483
+ "epoch": 1.5,
1484
+ "grad_norm": 2.417782783508301,
1485
+ "learning_rate": 5e-05,
1486
+ "loss": 0.263,
1487
+ "step": 2100
1488
+ },
1489
+ {
1490
+ "epoch": 1.5071428571428571,
1491
+ "grad_norm": 3.639488935470581,
1492
+ "learning_rate": 4.976190476190477e-05,
1493
+ "loss": 0.2855,
1494
+ "step": 2110
1495
+ },
1496
+ {
1497
+ "epoch": 1.5142857142857142,
1498
+ "grad_norm": 3.0365843772888184,
1499
+ "learning_rate": 4.9523809523809525e-05,
1500
+ "loss": 0.3073,
1501
+ "step": 2120
1502
+ },
1503
+ {
1504
+ "epoch": 1.5214285714285714,
1505
+ "grad_norm": 4.151659965515137,
1506
+ "learning_rate": 4.928571428571429e-05,
1507
+ "loss": 0.2918,
1508
+ "step": 2130
1509
+ },
1510
+ {
1511
+ "epoch": 1.5285714285714285,
1512
+ "grad_norm": 11.354039192199707,
1513
+ "learning_rate": 4.904761904761905e-05,
1514
+ "loss": 0.3428,
1515
+ "step": 2140
1516
+ },
1517
+ {
1518
+ "epoch": 1.5357142857142856,
1519
+ "grad_norm": 3.684797525405884,
1520
+ "learning_rate": 4.880952380952381e-05,
1521
+ "loss": 0.2791,
1522
+ "step": 2150
1523
+ },
1524
+ {
1525
+ "epoch": 1.5428571428571427,
1526
+ "grad_norm": 3.1418874263763428,
1527
+ "learning_rate": 4.8571428571428576e-05,
1528
+ "loss": 0.2386,
1529
+ "step": 2160
1530
+ },
1531
+ {
1532
+ "epoch": 1.55,
1533
+ "grad_norm": 2.303208827972412,
1534
+ "learning_rate": 4.8333333333333334e-05,
1535
+ "loss": 0.3338,
1536
+ "step": 2170
1537
+ },
1538
+ {
1539
+ "epoch": 1.5571428571428572,
1540
+ "grad_norm": 7.349155902862549,
1541
+ "learning_rate": 4.80952380952381e-05,
1542
+ "loss": 0.2595,
1543
+ "step": 2180
1544
+ },
1545
+ {
1546
+ "epoch": 1.5642857142857143,
1547
+ "grad_norm": 2.4980363845825195,
1548
+ "learning_rate": 4.785714285714286e-05,
1549
+ "loss": 0.2499,
1550
+ "step": 2190
1551
+ },
1552
+ {
1553
+ "epoch": 1.5714285714285714,
1554
+ "grad_norm": 3.0202884674072266,
1555
+ "learning_rate": 4.761904761904762e-05,
1556
+ "loss": 0.2608,
1557
+ "step": 2200
1558
+ },
1559
+ {
1560
+ "epoch": 1.5785714285714287,
1561
+ "grad_norm": 7.224872589111328,
1562
+ "learning_rate": 4.738095238095238e-05,
1563
+ "loss": 0.3299,
1564
+ "step": 2210
1565
+ },
1566
+ {
1567
+ "epoch": 1.5857142857142859,
1568
+ "grad_norm": 3.9374120235443115,
1569
+ "learning_rate": 4.714285714285714e-05,
1570
+ "loss": 0.3279,
1571
+ "step": 2220
1572
+ },
1573
+ {
1574
+ "epoch": 1.592857142857143,
1575
+ "grad_norm": 5.184910774230957,
1576
+ "learning_rate": 4.690476190476191e-05,
1577
+ "loss": 0.2691,
1578
+ "step": 2230
1579
+ },
1580
+ {
1581
+ "epoch": 1.6,
1582
+ "grad_norm": 5.877132415771484,
1583
+ "learning_rate": 4.666666666666667e-05,
1584
+ "loss": 0.2903,
1585
+ "step": 2240
1586
+ },
1587
+ {
1588
+ "epoch": 1.6071428571428572,
1589
+ "grad_norm": 3.351426601409912,
1590
+ "learning_rate": 4.642857142857143e-05,
1591
+ "loss": 0.2771,
1592
+ "step": 2250
1593
+ },
1594
+ {
1595
+ "epoch": 1.6142857142857143,
1596
+ "grad_norm": 1.7859790325164795,
1597
+ "learning_rate": 4.6190476190476194e-05,
1598
+ "loss": 0.349,
1599
+ "step": 2260
1600
+ },
1601
+ {
1602
+ "epoch": 1.6214285714285714,
1603
+ "grad_norm": 1.9868186712265015,
1604
+ "learning_rate": 4.595238095238095e-05,
1605
+ "loss": 0.3505,
1606
+ "step": 2270
1607
+ },
1608
+ {
1609
+ "epoch": 1.6285714285714286,
1610
+ "grad_norm": 3.075303316116333,
1611
+ "learning_rate": 4.5714285714285716e-05,
1612
+ "loss": 0.2753,
1613
+ "step": 2280
1614
+ },
1615
+ {
1616
+ "epoch": 1.6357142857142857,
1617
+ "grad_norm": 2.1696481704711914,
1618
+ "learning_rate": 4.547619047619048e-05,
1619
+ "loss": 0.2501,
1620
+ "step": 2290
1621
+ },
1622
+ {
1623
+ "epoch": 1.6428571428571428,
1624
+ "grad_norm": 2.9309706687927246,
1625
+ "learning_rate": 4.523809523809524e-05,
1626
+ "loss": 0.2904,
1627
+ "step": 2300
1628
+ },
1629
+ {
1630
+ "epoch": 1.65,
1631
+ "grad_norm": 3.5121254920959473,
1632
+ "learning_rate": 4.5e-05,
1633
+ "loss": 0.2786,
1634
+ "step": 2310
1635
+ },
1636
+ {
1637
+ "epoch": 1.657142857142857,
1638
+ "grad_norm": 1.3729180097579956,
1639
+ "learning_rate": 4.476190476190477e-05,
1640
+ "loss": 0.2589,
1641
+ "step": 2320
1642
+ },
1643
+ {
1644
+ "epoch": 1.6642857142857141,
1645
+ "grad_norm": 4.851173400878906,
1646
+ "learning_rate": 4.4523809523809525e-05,
1647
+ "loss": 0.3489,
1648
+ "step": 2330
1649
+ },
1650
+ {
1651
+ "epoch": 1.6714285714285713,
1652
+ "grad_norm": 4.376727104187012,
1653
+ "learning_rate": 4.428571428571428e-05,
1654
+ "loss": 0.3188,
1655
+ "step": 2340
1656
+ },
1657
+ {
1658
+ "epoch": 1.6785714285714286,
1659
+ "grad_norm": 2.563443183898926,
1660
+ "learning_rate": 4.404761904761905e-05,
1661
+ "loss": 0.3499,
1662
+ "step": 2350
1663
+ },
1664
+ {
1665
+ "epoch": 1.6857142857142857,
1666
+ "grad_norm": 2.768203020095825,
1667
+ "learning_rate": 4.380952380952381e-05,
1668
+ "loss": 0.2616,
1669
+ "step": 2360
1670
+ },
1671
+ {
1672
+ "epoch": 1.6928571428571428,
1673
+ "grad_norm": 4.015028476715088,
1674
+ "learning_rate": 4.3571428571428576e-05,
1675
+ "loss": 0.2851,
1676
+ "step": 2370
1677
+ },
1678
+ {
1679
+ "epoch": 1.7,
1680
+ "grad_norm": 1.4554102420806885,
1681
+ "learning_rate": 4.3333333333333334e-05,
1682
+ "loss": 0.274,
1683
+ "step": 2380
1684
+ },
1685
+ {
1686
+ "epoch": 1.7071428571428573,
1687
+ "grad_norm": 2.553905963897705,
1688
+ "learning_rate": 4.30952380952381e-05,
1689
+ "loss": 0.2692,
1690
+ "step": 2390
1691
+ },
1692
+ {
1693
+ "epoch": 1.7142857142857144,
1694
+ "grad_norm": 3.319626569747925,
1695
+ "learning_rate": 4.2857142857142856e-05,
1696
+ "loss": 0.3143,
1697
+ "step": 2400
1698
+ },
1699
+ {
1700
+ "epoch": 1.7214285714285715,
1701
+ "grad_norm": 2.207327127456665,
1702
+ "learning_rate": 4.261904761904762e-05,
1703
+ "loss": 0.2548,
1704
+ "step": 2410
1705
+ },
1706
+ {
1707
+ "epoch": 1.7285714285714286,
1708
+ "grad_norm": 3.1698238849639893,
1709
+ "learning_rate": 4.2380952380952385e-05,
1710
+ "loss": 0.3354,
1711
+ "step": 2420
1712
+ },
1713
+ {
1714
+ "epoch": 1.7357142857142858,
1715
+ "grad_norm": 5.368778228759766,
1716
+ "learning_rate": 4.214285714285714e-05,
1717
+ "loss": 0.2665,
1718
+ "step": 2430
1719
+ },
1720
+ {
1721
+ "epoch": 1.7428571428571429,
1722
+ "grad_norm": 1.1383056640625,
1723
+ "learning_rate": 4.190476190476191e-05,
1724
+ "loss": 0.2947,
1725
+ "step": 2440
1726
+ },
1727
+ {
1728
+ "epoch": 1.75,
1729
+ "grad_norm": 3.5954413414001465,
1730
+ "learning_rate": 4.166666666666667e-05,
1731
+ "loss": 0.3362,
1732
+ "step": 2450
1733
+ },
1734
+ {
1735
+ "epoch": 1.7571428571428571,
1736
+ "grad_norm": 3.8164236545562744,
1737
+ "learning_rate": 4.1428571428571437e-05,
1738
+ "loss": 0.2375,
1739
+ "step": 2460
1740
+ },
1741
+ {
1742
+ "epoch": 1.7642857142857142,
1743
+ "grad_norm": 4.327922344207764,
1744
+ "learning_rate": 4.119047619047619e-05,
1745
+ "loss": 0.2776,
1746
+ "step": 2470
1747
+ },
1748
+ {
1749
+ "epoch": 1.7714285714285714,
1750
+ "grad_norm": 2.2782583236694336,
1751
+ "learning_rate": 4.095238095238095e-05,
1752
+ "loss": 0.2407,
1753
+ "step": 2480
1754
+ },
1755
+ {
1756
+ "epoch": 1.7785714285714285,
1757
+ "grad_norm": 1.7074918746948242,
1758
+ "learning_rate": 4.0714285714285717e-05,
1759
+ "loss": 0.2514,
1760
+ "step": 2490
1761
+ },
1762
+ {
1763
+ "epoch": 1.7857142857142856,
1764
+ "grad_norm": 4.881974220275879,
1765
+ "learning_rate": 4.047619047619048e-05,
1766
+ "loss": 0.3229,
1767
+ "step": 2500
1768
+ },
1769
+ {
1770
+ "epoch": 1.7928571428571427,
1771
+ "grad_norm": 4.011612892150879,
1772
+ "learning_rate": 4.023809523809524e-05,
1773
+ "loss": 0.3093,
1774
+ "step": 2510
1775
+ },
1776
+ {
1777
+ "epoch": 1.8,
1778
+ "grad_norm": 2.8738279342651367,
1779
+ "learning_rate": 4e-05,
1780
+ "loss": 0.2844,
1781
+ "step": 2520
1782
+ },
1783
+ {
1784
+ "epoch": 1.8071428571428572,
1785
+ "grad_norm": 2.006197214126587,
1786
+ "learning_rate": 3.976190476190476e-05,
1787
+ "loss": 0.2299,
1788
+ "step": 2530
1789
+ },
1790
+ {
1791
+ "epoch": 1.8142857142857143,
1792
+ "grad_norm": 2.935206651687622,
1793
+ "learning_rate": 3.9523809523809526e-05,
1794
+ "loss": 0.2737,
1795
+ "step": 2540
1796
+ },
1797
+ {
1798
+ "epoch": 1.8214285714285714,
1799
+ "grad_norm": 2.227921962738037,
1800
+ "learning_rate": 3.928571428571429e-05,
1801
+ "loss": 0.2654,
1802
+ "step": 2550
1803
+ },
1804
+ {
1805
+ "epoch": 1.8285714285714287,
1806
+ "grad_norm": 5.1243367195129395,
1807
+ "learning_rate": 3.904761904761905e-05,
1808
+ "loss": 0.3842,
1809
+ "step": 2560
1810
+ },
1811
+ {
1812
+ "epoch": 1.8357142857142859,
1813
+ "grad_norm": 2.0849359035491943,
1814
+ "learning_rate": 3.880952380952381e-05,
1815
+ "loss": 0.2521,
1816
+ "step": 2570
1817
+ },
1818
+ {
1819
+ "epoch": 1.842857142857143,
1820
+ "grad_norm": 2.884455442428589,
1821
+ "learning_rate": 3.857142857142858e-05,
1822
+ "loss": 0.2784,
1823
+ "step": 2580
1824
+ },
1825
+ {
1826
+ "epoch": 1.85,
1827
+ "grad_norm": 2.0989837646484375,
1828
+ "learning_rate": 3.8333333333333334e-05,
1829
+ "loss": 0.2359,
1830
+ "step": 2590
1831
+ },
1832
+ {
1833
+ "epoch": 1.8571428571428572,
1834
+ "grad_norm": 3.074004650115967,
1835
+ "learning_rate": 3.809523809523809e-05,
1836
+ "loss": 0.2629,
1837
+ "step": 2600
1838
+ },
1839
+ {
1840
+ "epoch": 1.8642857142857143,
1841
+ "grad_norm": 3.4337799549102783,
1842
+ "learning_rate": 3.785714285714286e-05,
1843
+ "loss": 0.2449,
1844
+ "step": 2610
1845
+ },
1846
+ {
1847
+ "epoch": 1.8714285714285714,
1848
+ "grad_norm": 4.196152687072754,
1849
+ "learning_rate": 3.761904761904762e-05,
1850
+ "loss": 0.3082,
1851
+ "step": 2620
1852
+ },
1853
+ {
1854
+ "epoch": 1.8785714285714286,
1855
+ "grad_norm": 2.139364719390869,
1856
+ "learning_rate": 3.7380952380952386e-05,
1857
+ "loss": 0.2816,
1858
+ "step": 2630
1859
+ },
1860
+ {
1861
+ "epoch": 1.8857142857142857,
1862
+ "grad_norm": 6.65344762802124,
1863
+ "learning_rate": 3.7142857142857143e-05,
1864
+ "loss": 0.2729,
1865
+ "step": 2640
1866
+ },
1867
+ {
1868
+ "epoch": 1.8928571428571428,
1869
+ "grad_norm": 1.8867508172988892,
1870
+ "learning_rate": 3.690476190476191e-05,
1871
+ "loss": 0.2668,
1872
+ "step": 2650
1873
+ },
1874
+ {
1875
+ "epoch": 1.9,
1876
+ "grad_norm": 4.062267303466797,
1877
+ "learning_rate": 3.6666666666666666e-05,
1878
+ "loss": 0.2908,
1879
+ "step": 2660
1880
+ },
1881
+ {
1882
+ "epoch": 1.907142857142857,
1883
+ "grad_norm": 4.145746231079102,
1884
+ "learning_rate": 3.642857142857143e-05,
1885
+ "loss": 0.3168,
1886
+ "step": 2670
1887
+ },
1888
+ {
1889
+ "epoch": 1.9142857142857141,
1890
+ "grad_norm": 6.509150505065918,
1891
+ "learning_rate": 3.619047619047619e-05,
1892
+ "loss": 0.2744,
1893
+ "step": 2680
1894
+ },
1895
+ {
1896
+ "epoch": 1.9214285714285713,
1897
+ "grad_norm": 2.091609239578247,
1898
+ "learning_rate": 3.595238095238095e-05,
1899
+ "loss": 0.2945,
1900
+ "step": 2690
1901
+ },
1902
+ {
1903
+ "epoch": 1.9285714285714286,
1904
+ "grad_norm": 2.6082661151885986,
1905
+ "learning_rate": 3.571428571428572e-05,
1906
+ "loss": 0.2538,
1907
+ "step": 2700
1908
+ },
1909
+ {
1910
+ "epoch": 1.9357142857142857,
1911
+ "grad_norm": 3.4130637645721436,
1912
+ "learning_rate": 3.547619047619048e-05,
1913
+ "loss": 0.2726,
1914
+ "step": 2710
1915
+ },
1916
+ {
1917
+ "epoch": 1.9428571428571428,
1918
+ "grad_norm": 5.003239631652832,
1919
+ "learning_rate": 3.523809523809524e-05,
1920
+ "loss": 0.2697,
1921
+ "step": 2720
1922
+ },
1923
+ {
1924
+ "epoch": 1.95,
1925
+ "grad_norm": 2.7061688899993896,
1926
+ "learning_rate": 3.5e-05,
1927
+ "loss": 0.2756,
1928
+ "step": 2730
1929
+ },
1930
+ {
1931
+ "epoch": 1.9571428571428573,
1932
+ "grad_norm": 2.0089480876922607,
1933
+ "learning_rate": 3.476190476190476e-05,
1934
+ "loss": 0.2869,
1935
+ "step": 2740
1936
+ },
1937
+ {
1938
+ "epoch": 1.9642857142857144,
1939
+ "grad_norm": 2.2568588256835938,
1940
+ "learning_rate": 3.4523809523809526e-05,
1941
+ "loss": 0.2687,
1942
+ "step": 2750
1943
+ },
1944
+ {
1945
+ "epoch": 1.9714285714285715,
1946
+ "grad_norm": 2.0823915004730225,
1947
+ "learning_rate": 3.428571428571429e-05,
1948
+ "loss": 0.263,
1949
+ "step": 2760
1950
+ },
1951
+ {
1952
+ "epoch": 1.9785714285714286,
1953
+ "grad_norm": 1.738771915435791,
1954
+ "learning_rate": 3.404761904761905e-05,
1955
+ "loss": 0.2283,
1956
+ "step": 2770
1957
+ },
1958
+ {
1959
+ "epoch": 1.9857142857142858,
1960
+ "grad_norm": 4.0020623207092285,
1961
+ "learning_rate": 3.380952380952381e-05,
1962
+ "loss": 0.3286,
1963
+ "step": 2780
1964
+ },
1965
+ {
1966
+ "epoch": 1.9928571428571429,
1967
+ "grad_norm": 2.4708666801452637,
1968
+ "learning_rate": 3.357142857142857e-05,
1969
+ "loss": 0.279,
1970
+ "step": 2790
1971
+ },
1972
+ {
1973
+ "epoch": 2.0,
1974
+ "grad_norm": 1.950387716293335,
1975
+ "learning_rate": 3.3333333333333335e-05,
1976
+ "loss": 0.3078,
1977
+ "step": 2800
1978
+ },
1979
+ {
1980
+ "epoch": 2.0,
1981
+ "eval_loss": 0.25949418544769287,
1982
+ "eval_runtime": 68.2677,
1983
+ "eval_samples_per_second": 2.93,
1984
+ "eval_steps_per_second": 0.366,
1985
+ "step": 2800
1986
+ },
1987
+ {
1988
+ "epoch": 2.007142857142857,
1989
+ "grad_norm": 2.9532036781311035,
1990
+ "learning_rate": 3.309523809523809e-05,
1991
+ "loss": 0.2379,
1992
+ "step": 2810
1993
+ },
1994
+ {
1995
+ "epoch": 2.0142857142857142,
1996
+ "grad_norm": 2.3207550048828125,
1997
+ "learning_rate": 3.285714285714286e-05,
1998
+ "loss": 0.3187,
1999
+ "step": 2820
2000
+ },
2001
+ {
2002
+ "epoch": 2.0214285714285714,
2003
+ "grad_norm": 1.4604542255401611,
2004
+ "learning_rate": 3.261904761904762e-05,
2005
+ "loss": 0.2192,
2006
+ "step": 2830
2007
+ },
2008
+ {
2009
+ "epoch": 2.0285714285714285,
2010
+ "grad_norm": 2.958681106567383,
2011
+ "learning_rate": 3.2380952380952386e-05,
2012
+ "loss": 0.2237,
2013
+ "step": 2840
2014
+ },
2015
+ {
2016
+ "epoch": 2.0357142857142856,
2017
+ "grad_norm": 4.185633182525635,
2018
+ "learning_rate": 3.2142857142857144e-05,
2019
+ "loss": 0.3259,
2020
+ "step": 2850
2021
+ },
2022
+ {
2023
+ "epoch": 2.0428571428571427,
2024
+ "grad_norm": 3.2991695404052734,
2025
+ "learning_rate": 3.19047619047619e-05,
2026
+ "loss": 0.1868,
2027
+ "step": 2860
2028
+ },
2029
+ {
2030
+ "epoch": 2.05,
2031
+ "grad_norm": 2.3345556259155273,
2032
+ "learning_rate": 3.1666666666666666e-05,
2033
+ "loss": 0.2627,
2034
+ "step": 2870
2035
+ },
2036
+ {
2037
+ "epoch": 2.057142857142857,
2038
+ "grad_norm": 1.6416515111923218,
2039
+ "learning_rate": 3.142857142857143e-05,
2040
+ "loss": 0.2417,
2041
+ "step": 2880
2042
+ },
2043
+ {
2044
+ "epoch": 2.064285714285714,
2045
+ "grad_norm": 2.0688915252685547,
2046
+ "learning_rate": 3.1190476190476195e-05,
2047
+ "loss": 0.2359,
2048
+ "step": 2890
2049
+ },
2050
+ {
2051
+ "epoch": 2.0714285714285716,
2052
+ "grad_norm": 2.668055534362793,
2053
+ "learning_rate": 3.095238095238095e-05,
2054
+ "loss": 0.2879,
2055
+ "step": 2900
2056
+ },
2057
+ {
2058
+ "epoch": 2.0785714285714287,
2059
+ "grad_norm": 2.9602174758911133,
2060
+ "learning_rate": 3.071428571428572e-05,
2061
+ "loss": 0.2061,
2062
+ "step": 2910
2063
+ },
2064
+ {
2065
+ "epoch": 2.085714285714286,
2066
+ "grad_norm": 3.5890145301818848,
2067
+ "learning_rate": 3.0476190476190482e-05,
2068
+ "loss": 0.266,
2069
+ "step": 2920
2070
+ },
2071
+ {
2072
+ "epoch": 2.092857142857143,
2073
+ "grad_norm": 2.9162445068359375,
2074
+ "learning_rate": 3.0238095238095236e-05,
2075
+ "loss": 0.234,
2076
+ "step": 2930
2077
+ },
2078
+ {
2079
+ "epoch": 2.1,
2080
+ "grad_norm": 5.290048599243164,
2081
+ "learning_rate": 3e-05,
2082
+ "loss": 0.2562,
2083
+ "step": 2940
2084
+ },
2085
+ {
2086
+ "epoch": 2.107142857142857,
2087
+ "grad_norm": 4.895926475524902,
2088
+ "learning_rate": 2.9761904761904762e-05,
2089
+ "loss": 0.2958,
2090
+ "step": 2950
2091
+ },
2092
+ {
2093
+ "epoch": 2.1142857142857143,
2094
+ "grad_norm": 5.531802177429199,
2095
+ "learning_rate": 2.9523809523809526e-05,
2096
+ "loss": 0.2388,
2097
+ "step": 2960
2098
+ },
2099
+ {
2100
+ "epoch": 2.1214285714285714,
2101
+ "grad_norm": 2.0870862007141113,
2102
+ "learning_rate": 2.9285714285714288e-05,
2103
+ "loss": 0.3,
2104
+ "step": 2970
2105
+ },
2106
+ {
2107
+ "epoch": 2.1285714285714286,
2108
+ "grad_norm": 1.2288978099822998,
2109
+ "learning_rate": 2.9047619047619052e-05,
2110
+ "loss": 0.2691,
2111
+ "step": 2980
2112
+ },
2113
+ {
2114
+ "epoch": 2.1357142857142857,
2115
+ "grad_norm": 3.008167028427124,
2116
+ "learning_rate": 2.880952380952381e-05,
2117
+ "loss": 0.3721,
2118
+ "step": 2990
2119
+ },
2120
+ {
2121
+ "epoch": 2.142857142857143,
2122
+ "grad_norm": 4.923089027404785,
2123
+ "learning_rate": 2.857142857142857e-05,
2124
+ "loss": 0.2516,
2125
+ "step": 3000
2126
+ },
2127
+ {
2128
+ "epoch": 2.15,
2129
+ "grad_norm": 3.881120443344116,
2130
+ "learning_rate": 2.8333333333333335e-05,
2131
+ "loss": 0.275,
2132
+ "step": 3010
2133
+ },
2134
+ {
2135
+ "epoch": 2.157142857142857,
2136
+ "grad_norm": 4.30492639541626,
2137
+ "learning_rate": 2.8095238095238096e-05,
2138
+ "loss": 0.2353,
2139
+ "step": 3020
2140
+ },
2141
+ {
2142
+ "epoch": 2.164285714285714,
2143
+ "grad_norm": 3.907482862472534,
2144
+ "learning_rate": 2.785714285714286e-05,
2145
+ "loss": 0.2856,
2146
+ "step": 3030
2147
+ },
2148
+ {
2149
+ "epoch": 2.1714285714285713,
2150
+ "grad_norm": 4.306923866271973,
2151
+ "learning_rate": 2.7619047619047622e-05,
2152
+ "loss": 0.2768,
2153
+ "step": 3040
2154
+ },
2155
+ {
2156
+ "epoch": 2.1785714285714284,
2157
+ "grad_norm": 4.364368915557861,
2158
+ "learning_rate": 2.7380952380952383e-05,
2159
+ "loss": 0.2854,
2160
+ "step": 3050
2161
+ },
2162
+ {
2163
+ "epoch": 2.185714285714286,
2164
+ "grad_norm": 3.7898383140563965,
2165
+ "learning_rate": 2.714285714285714e-05,
2166
+ "loss": 0.2168,
2167
+ "step": 3060
2168
+ },
2169
+ {
2170
+ "epoch": 2.192857142857143,
2171
+ "grad_norm": 5.487813472747803,
2172
+ "learning_rate": 2.6904761904761905e-05,
2173
+ "loss": 0.2849,
2174
+ "step": 3070
2175
+ },
2176
+ {
2177
+ "epoch": 2.2,
2178
+ "grad_norm": 3.3784303665161133,
2179
+ "learning_rate": 2.6666666666666667e-05,
2180
+ "loss": 0.2431,
2181
+ "step": 3080
2182
+ },
2183
+ {
2184
+ "epoch": 2.2071428571428573,
2185
+ "grad_norm": 4.644072532653809,
2186
+ "learning_rate": 2.642857142857143e-05,
2187
+ "loss": 0.3144,
2188
+ "step": 3090
2189
+ },
2190
+ {
2191
+ "epoch": 2.2142857142857144,
2192
+ "grad_norm": 9.652532577514648,
2193
+ "learning_rate": 2.6190476190476192e-05,
2194
+ "loss": 0.3005,
2195
+ "step": 3100
2196
+ },
2197
+ {
2198
+ "epoch": 2.2214285714285715,
2199
+ "grad_norm": 2.1053450107574463,
2200
+ "learning_rate": 2.5952380952380957e-05,
2201
+ "loss": 0.2487,
2202
+ "step": 3110
2203
+ },
2204
+ {
2205
+ "epoch": 2.2285714285714286,
2206
+ "grad_norm": 3.275733709335327,
2207
+ "learning_rate": 2.5714285714285714e-05,
2208
+ "loss": 0.2863,
2209
+ "step": 3120
2210
+ },
2211
+ {
2212
+ "epoch": 2.2357142857142858,
2213
+ "grad_norm": 2.46889066696167,
2214
+ "learning_rate": 2.5476190476190476e-05,
2215
+ "loss": 0.2497,
2216
+ "step": 3130
2217
+ },
2218
+ {
2219
+ "epoch": 2.242857142857143,
2220
+ "grad_norm": 4.135228633880615,
2221
+ "learning_rate": 2.523809523809524e-05,
2222
+ "loss": 0.2368,
2223
+ "step": 3140
2224
+ },
2225
+ {
2226
+ "epoch": 2.25,
2227
+ "grad_norm": 5.32862663269043,
2228
+ "learning_rate": 2.5e-05,
2229
+ "loss": 0.2336,
2230
+ "step": 3150
2231
+ },
2232
+ {
2233
+ "epoch": 2.257142857142857,
2234
+ "grad_norm": 5.055444717407227,
2235
+ "learning_rate": 2.4761904761904762e-05,
2236
+ "loss": 0.2474,
2237
+ "step": 3160
2238
+ },
2239
+ {
2240
+ "epoch": 2.2642857142857142,
2241
+ "grad_norm": 9.018756866455078,
2242
+ "learning_rate": 2.4523809523809523e-05,
2243
+ "loss": 0.2637,
2244
+ "step": 3170
2245
+ },
2246
+ {
2247
+ "epoch": 2.2714285714285714,
2248
+ "grad_norm": 2.1222236156463623,
2249
+ "learning_rate": 2.4285714285714288e-05,
2250
+ "loss": 0.2242,
2251
+ "step": 3180
2252
+ },
2253
+ {
2254
+ "epoch": 2.2785714285714285,
2255
+ "grad_norm": 5.100944519042969,
2256
+ "learning_rate": 2.404761904761905e-05,
2257
+ "loss": 0.2634,
2258
+ "step": 3190
2259
+ },
2260
+ {
2261
+ "epoch": 2.2857142857142856,
2262
+ "grad_norm": 5.248310565948486,
2263
+ "learning_rate": 2.380952380952381e-05,
2264
+ "loss": 0.2557,
2265
+ "step": 3200
2266
+ },
2267
+ {
2268
+ "epoch": 2.2928571428571427,
2269
+ "grad_norm": 4.782702922821045,
2270
+ "learning_rate": 2.357142857142857e-05,
2271
+ "loss": 0.2313,
2272
+ "step": 3210
2273
+ },
2274
+ {
2275
+ "epoch": 2.3,
2276
+ "grad_norm": 5.026177883148193,
2277
+ "learning_rate": 2.3333333333333336e-05,
2278
+ "loss": 0.3054,
2279
+ "step": 3220
2280
+ },
2281
+ {
2282
+ "epoch": 2.307142857142857,
2283
+ "grad_norm": 2.8892881870269775,
2284
+ "learning_rate": 2.3095238095238097e-05,
2285
+ "loss": 0.2477,
2286
+ "step": 3230
2287
+ },
2288
+ {
2289
+ "epoch": 2.314285714285714,
2290
+ "grad_norm": 3.408980131149292,
2291
+ "learning_rate": 2.2857142857142858e-05,
2292
+ "loss": 0.2381,
2293
+ "step": 3240
2294
+ },
2295
+ {
2296
+ "epoch": 2.3214285714285716,
2297
+ "grad_norm": 2.3508706092834473,
2298
+ "learning_rate": 2.261904761904762e-05,
2299
+ "loss": 0.1975,
2300
+ "step": 3250
2301
+ },
2302
+ {
2303
+ "epoch": 2.3285714285714287,
2304
+ "grad_norm": 1.3427135944366455,
2305
+ "learning_rate": 2.2380952380952384e-05,
2306
+ "loss": 0.2334,
2307
+ "step": 3260
2308
+ },
2309
+ {
2310
+ "epoch": 2.335714285714286,
2311
+ "grad_norm": 3.8124921321868896,
2312
+ "learning_rate": 2.214285714285714e-05,
2313
+ "loss": 0.2983,
2314
+ "step": 3270
2315
+ },
2316
+ {
2317
+ "epoch": 2.342857142857143,
2318
+ "grad_norm": 3.923583507537842,
2319
+ "learning_rate": 2.1904761904761906e-05,
2320
+ "loss": 0.278,
2321
+ "step": 3280
2322
+ },
2323
+ {
2324
+ "epoch": 2.35,
2325
+ "grad_norm": 3.14457106590271,
2326
+ "learning_rate": 2.1666666666666667e-05,
2327
+ "loss": 0.2409,
2328
+ "step": 3290
2329
+ },
2330
+ {
2331
+ "epoch": 2.357142857142857,
2332
+ "grad_norm": 4.7329230308532715,
2333
+ "learning_rate": 2.1428571428571428e-05,
2334
+ "loss": 0.2861,
2335
+ "step": 3300
2336
+ },
2337
+ {
2338
+ "epoch": 2.3642857142857143,
2339
+ "grad_norm": 5.68926477432251,
2340
+ "learning_rate": 2.1190476190476193e-05,
2341
+ "loss": 0.2904,
2342
+ "step": 3310
2343
+ },
2344
+ {
2345
+ "epoch": 2.3714285714285714,
2346
+ "grad_norm": 8.324594497680664,
2347
+ "learning_rate": 2.0952380952380954e-05,
2348
+ "loss": 0.3381,
2349
+ "step": 3320
2350
+ },
2351
+ {
2352
+ "epoch": 2.3785714285714286,
2353
+ "grad_norm": 5.043862819671631,
2354
+ "learning_rate": 2.0714285714285718e-05,
2355
+ "loss": 0.2893,
2356
+ "step": 3330
2357
+ },
2358
+ {
2359
+ "epoch": 2.3857142857142857,
2360
+ "grad_norm": 1.9703952074050903,
2361
+ "learning_rate": 2.0476190476190476e-05,
2362
+ "loss": 0.1997,
2363
+ "step": 3340
2364
+ },
2365
+ {
2366
+ "epoch": 2.392857142857143,
2367
+ "grad_norm": 3.2141919136047363,
2368
+ "learning_rate": 2.023809523809524e-05,
2369
+ "loss": 0.2547,
2370
+ "step": 3350
2371
+ },
2372
+ {
2373
+ "epoch": 2.4,
2374
+ "grad_norm": 3.0841588973999023,
2375
+ "learning_rate": 2e-05,
2376
+ "loss": 0.2527,
2377
+ "step": 3360
2378
+ },
2379
+ {
2380
+ "epoch": 2.407142857142857,
2381
+ "grad_norm": 2.329394817352295,
2382
+ "learning_rate": 1.9761904761904763e-05,
2383
+ "loss": 0.2228,
2384
+ "step": 3370
2385
+ },
2386
+ {
2387
+ "epoch": 2.414285714285714,
2388
+ "grad_norm": 2.290874719619751,
2389
+ "learning_rate": 1.9523809523809524e-05,
2390
+ "loss": 0.2613,
2391
+ "step": 3380
2392
+ },
2393
+ {
2394
+ "epoch": 2.4214285714285713,
2395
+ "grad_norm": 1.3011747598648071,
2396
+ "learning_rate": 1.928571428571429e-05,
2397
+ "loss": 0.22,
2398
+ "step": 3390
2399
+ },
2400
+ {
2401
+ "epoch": 2.4285714285714284,
2402
+ "grad_norm": 3.5783472061157227,
2403
+ "learning_rate": 1.9047619047619046e-05,
2404
+ "loss": 0.2729,
2405
+ "step": 3400
2406
+ },
2407
+ {
2408
+ "epoch": 2.435714285714286,
2409
+ "grad_norm": 3.1245083808898926,
2410
+ "learning_rate": 1.880952380952381e-05,
2411
+ "loss": 0.2551,
2412
+ "step": 3410
2413
+ },
2414
+ {
2415
+ "epoch": 2.442857142857143,
2416
+ "grad_norm": 4.986431121826172,
2417
+ "learning_rate": 1.8571428571428572e-05,
2418
+ "loss": 0.3029,
2419
+ "step": 3420
2420
+ },
2421
+ {
2422
+ "epoch": 2.45,
2423
+ "grad_norm": 1.8483643531799316,
2424
+ "learning_rate": 1.8333333333333333e-05,
2425
+ "loss": 0.2458,
2426
+ "step": 3430
2427
+ },
2428
+ {
2429
+ "epoch": 2.4571428571428573,
2430
+ "grad_norm": 5.947587490081787,
2431
+ "learning_rate": 1.8095238095238094e-05,
2432
+ "loss": 0.2848,
2433
+ "step": 3440
2434
+ },
2435
+ {
2436
+ "epoch": 2.4642857142857144,
2437
+ "grad_norm": 5.204590320587158,
2438
+ "learning_rate": 1.785714285714286e-05,
2439
+ "loss": 0.3195,
2440
+ "step": 3450
2441
+ },
2442
+ {
2443
+ "epoch": 2.4714285714285715,
2444
+ "grad_norm": 2.213702440261841,
2445
+ "learning_rate": 1.761904761904762e-05,
2446
+ "loss": 0.2624,
2447
+ "step": 3460
2448
+ },
2449
+ {
2450
+ "epoch": 2.4785714285714286,
2451
+ "grad_norm": 2.896759033203125,
2452
+ "learning_rate": 1.738095238095238e-05,
2453
+ "loss": 0.3234,
2454
+ "step": 3470
2455
+ },
2456
+ {
2457
+ "epoch": 2.4857142857142858,
2458
+ "grad_norm": 2.1793391704559326,
2459
+ "learning_rate": 1.7142857142857145e-05,
2460
+ "loss": 0.2136,
2461
+ "step": 3480
2462
+ },
2463
+ {
2464
+ "epoch": 2.492857142857143,
2465
+ "grad_norm": 5.157558917999268,
2466
+ "learning_rate": 1.6904761904761906e-05,
2467
+ "loss": 0.2655,
2468
+ "step": 3490
2469
+ },
2470
+ {
2471
+ "epoch": 2.5,
2472
+ "grad_norm": 2.9449174404144287,
2473
+ "learning_rate": 1.6666666666666667e-05,
2474
+ "loss": 0.2846,
2475
+ "step": 3500
2476
+ },
2477
+ {
2478
+ "epoch": 2.507142857142857,
2479
+ "grad_norm": 3.97674298286438,
2480
+ "learning_rate": 1.642857142857143e-05,
2481
+ "loss": 0.2423,
2482
+ "step": 3510
2483
+ },
2484
+ {
2485
+ "epoch": 2.5142857142857142,
2486
+ "grad_norm": 5.1696553230285645,
2487
+ "learning_rate": 1.6190476190476193e-05,
2488
+ "loss": 0.2468,
2489
+ "step": 3520
2490
+ },
2491
+ {
2492
+ "epoch": 2.5214285714285714,
2493
+ "grad_norm": 2.9929051399230957,
2494
+ "learning_rate": 1.595238095238095e-05,
2495
+ "loss": 0.2912,
2496
+ "step": 3530
2497
+ },
2498
+ {
2499
+ "epoch": 2.5285714285714285,
2500
+ "grad_norm": 3.287910223007202,
2501
+ "learning_rate": 1.5714285714285715e-05,
2502
+ "loss": 0.2436,
2503
+ "step": 3540
2504
+ },
2505
+ {
2506
+ "epoch": 2.5357142857142856,
2507
+ "grad_norm": 4.533077239990234,
2508
+ "learning_rate": 1.5476190476190476e-05,
2509
+ "loss": 0.3301,
2510
+ "step": 3550
2511
+ },
2512
+ {
2513
+ "epoch": 2.5428571428571427,
2514
+ "grad_norm": 4.241918087005615,
2515
+ "learning_rate": 1.5238095238095241e-05,
2516
+ "loss": 0.3111,
2517
+ "step": 3560
2518
+ },
2519
+ {
2520
+ "epoch": 2.55,
2521
+ "grad_norm": 2.0946638584136963,
2522
+ "learning_rate": 1.5e-05,
2523
+ "loss": 0.2098,
2524
+ "step": 3570
2525
+ },
2526
+ {
2527
+ "epoch": 2.557142857142857,
2528
+ "grad_norm": 4.646461486816406,
2529
+ "learning_rate": 1.4761904761904763e-05,
2530
+ "loss": 0.1745,
2531
+ "step": 3580
2532
+ },
2533
+ {
2534
+ "epoch": 2.564285714285714,
2535
+ "grad_norm": 4.6408233642578125,
2536
+ "learning_rate": 1.4523809523809526e-05,
2537
+ "loss": 0.212,
2538
+ "step": 3590
2539
+ },
2540
+ {
2541
+ "epoch": 2.571428571428571,
2542
+ "grad_norm": 6.365333080291748,
2543
+ "learning_rate": 1.4285714285714285e-05,
2544
+ "loss": 0.2445,
2545
+ "step": 3600
2546
+ },
2547
+ {
2548
+ "epoch": 2.5785714285714287,
2549
+ "grad_norm": 5.597777366638184,
2550
+ "learning_rate": 1.4047619047619048e-05,
2551
+ "loss": 0.2973,
2552
+ "step": 3610
2553
+ },
2554
+ {
2555
+ "epoch": 2.585714285714286,
2556
+ "grad_norm": 4.016028881072998,
2557
+ "learning_rate": 1.3809523809523811e-05,
2558
+ "loss": 0.2514,
2559
+ "step": 3620
2560
+ },
2561
+ {
2562
+ "epoch": 2.592857142857143,
2563
+ "grad_norm": 3.058223009109497,
2564
+ "learning_rate": 1.357142857142857e-05,
2565
+ "loss": 0.2766,
2566
+ "step": 3630
2567
+ },
2568
+ {
2569
+ "epoch": 2.6,
2570
+ "grad_norm": 2.698791027069092,
2571
+ "learning_rate": 1.3333333333333333e-05,
2572
+ "loss": 0.2042,
2573
+ "step": 3640
2574
+ },
2575
+ {
2576
+ "epoch": 2.607142857142857,
2577
+ "grad_norm": 5.222252368927002,
2578
+ "learning_rate": 1.3095238095238096e-05,
2579
+ "loss": 0.2953,
2580
+ "step": 3650
2581
+ },
2582
+ {
2583
+ "epoch": 2.6142857142857143,
2584
+ "grad_norm": 1.2256240844726562,
2585
+ "learning_rate": 1.2857142857142857e-05,
2586
+ "loss": 0.2938,
2587
+ "step": 3660
2588
+ },
2589
+ {
2590
+ "epoch": 2.6214285714285714,
2591
+ "grad_norm": 2.3339574337005615,
2592
+ "learning_rate": 1.261904761904762e-05,
2593
+ "loss": 0.2417,
2594
+ "step": 3670
2595
+ },
2596
+ {
2597
+ "epoch": 2.6285714285714286,
2598
+ "grad_norm": 6.824151039123535,
2599
+ "learning_rate": 1.2380952380952381e-05,
2600
+ "loss": 0.259,
2601
+ "step": 3680
2602
+ },
2603
+ {
2604
+ "epoch": 2.6357142857142857,
2605
+ "grad_norm": 5.956028461456299,
2606
+ "learning_rate": 1.2142857142857144e-05,
2607
+ "loss": 0.2088,
2608
+ "step": 3690
2609
+ },
2610
+ {
2611
+ "epoch": 2.642857142857143,
2612
+ "grad_norm": 1.825620174407959,
2613
+ "learning_rate": 1.1904761904761905e-05,
2614
+ "loss": 0.2426,
2615
+ "step": 3700
2616
+ },
2617
+ {
2618
+ "epoch": 2.65,
2619
+ "grad_norm": 1.8482383489608765,
2620
+ "learning_rate": 1.1666666666666668e-05,
2621
+ "loss": 0.2352,
2622
+ "step": 3710
2623
+ },
2624
+ {
2625
+ "epoch": 2.657142857142857,
2626
+ "grad_norm": 5.8184919357299805,
2627
+ "learning_rate": 1.1428571428571429e-05,
2628
+ "loss": 0.2556,
2629
+ "step": 3720
2630
+ },
2631
+ {
2632
+ "epoch": 2.664285714285714,
2633
+ "grad_norm": 3.534730911254883,
2634
+ "learning_rate": 1.1190476190476192e-05,
2635
+ "loss": 0.3155,
2636
+ "step": 3730
2637
+ },
2638
+ {
2639
+ "epoch": 2.6714285714285713,
2640
+ "grad_norm": 1.4020763635635376,
2641
+ "learning_rate": 1.0952380952380953e-05,
2642
+ "loss": 0.21,
2643
+ "step": 3740
2644
+ },
2645
+ {
2646
+ "epoch": 2.678571428571429,
2647
+ "grad_norm": 3.11566162109375,
2648
+ "learning_rate": 1.0714285714285714e-05,
2649
+ "loss": 0.2764,
2650
+ "step": 3750
2651
+ },
2652
+ {
2653
+ "epoch": 2.685714285714286,
2654
+ "grad_norm": 4.930715084075928,
2655
+ "learning_rate": 1.0476190476190477e-05,
2656
+ "loss": 0.3339,
2657
+ "step": 3760
2658
+ },
2659
+ {
2660
+ "epoch": 2.692857142857143,
2661
+ "grad_norm": 4.198614597320557,
2662
+ "learning_rate": 1.0238095238095238e-05,
2663
+ "loss": 0.2917,
2664
+ "step": 3770
2665
+ },
2666
+ {
2667
+ "epoch": 2.7,
2668
+ "grad_norm": 1.6880650520324707,
2669
+ "learning_rate": 1e-05,
2670
+ "loss": 0.2212,
2671
+ "step": 3780
2672
+ },
2673
+ {
2674
+ "epoch": 2.7071428571428573,
2675
+ "grad_norm": 3.580629587173462,
2676
+ "learning_rate": 9.761904761904762e-06,
2677
+ "loss": 0.2547,
2678
+ "step": 3790
2679
+ },
2680
+ {
2681
+ "epoch": 2.7142857142857144,
2682
+ "grad_norm": 2.764923095703125,
2683
+ "learning_rate": 9.523809523809523e-06,
2684
+ "loss": 0.2286,
2685
+ "step": 3800
2686
+ },
2687
+ {
2688
+ "epoch": 2.7214285714285715,
2689
+ "grad_norm": 2.3037569522857666,
2690
+ "learning_rate": 9.285714285714286e-06,
2691
+ "loss": 0.1979,
2692
+ "step": 3810
2693
+ },
2694
+ {
2695
+ "epoch": 2.7285714285714286,
2696
+ "grad_norm": 3.4528393745422363,
2697
+ "learning_rate": 9.047619047619047e-06,
2698
+ "loss": 0.2377,
2699
+ "step": 3820
2700
+ },
2701
+ {
2702
+ "epoch": 2.7357142857142858,
2703
+ "grad_norm": 2.326822519302368,
2704
+ "learning_rate": 8.80952380952381e-06,
2705
+ "loss": 0.2457,
2706
+ "step": 3830
2707
+ },
2708
+ {
2709
+ "epoch": 2.742857142857143,
2710
+ "grad_norm": 3.2105257511138916,
2711
+ "learning_rate": 8.571428571428573e-06,
2712
+ "loss": 0.2553,
2713
+ "step": 3840
2714
+ },
2715
+ {
2716
+ "epoch": 2.75,
2717
+ "grad_norm": 7.470506191253662,
2718
+ "learning_rate": 8.333333333333334e-06,
2719
+ "loss": 0.2285,
2720
+ "step": 3850
2721
+ },
2722
+ {
2723
+ "epoch": 2.757142857142857,
2724
+ "grad_norm": 7.341104984283447,
2725
+ "learning_rate": 8.095238095238097e-06,
2726
+ "loss": 0.2454,
2727
+ "step": 3860
2728
+ },
2729
+ {
2730
+ "epoch": 2.7642857142857142,
2731
+ "grad_norm": 5.3889055252075195,
2732
+ "learning_rate": 7.857142857142858e-06,
2733
+ "loss": 0.2754,
2734
+ "step": 3870
2735
+ },
2736
+ {
2737
+ "epoch": 2.7714285714285714,
2738
+ "grad_norm": 4.78093957901001,
2739
+ "learning_rate": 7.6190476190476205e-06,
2740
+ "loss": 0.3796,
2741
+ "step": 3880
2742
+ },
2743
+ {
2744
+ "epoch": 2.7785714285714285,
2745
+ "grad_norm": 2.088247060775757,
2746
+ "learning_rate": 7.380952380952382e-06,
2747
+ "loss": 0.2137,
2748
+ "step": 3890
2749
+ },
2750
+ {
2751
+ "epoch": 2.7857142857142856,
2752
+ "grad_norm": 2.5746066570281982,
2753
+ "learning_rate": 7.142857142857143e-06,
2754
+ "loss": 0.3069,
2755
+ "step": 3900
2756
+ },
2757
+ {
2758
+ "epoch": 2.7928571428571427,
2759
+ "grad_norm": 1.7371402978897095,
2760
+ "learning_rate": 6.9047619047619055e-06,
2761
+ "loss": 0.2522,
2762
+ "step": 3910
2763
+ },
2764
+ {
2765
+ "epoch": 2.8,
2766
+ "grad_norm": 1.5857880115509033,
2767
+ "learning_rate": 6.666666666666667e-06,
2768
+ "loss": 0.3117,
2769
+ "step": 3920
2770
+ },
2771
+ {
2772
+ "epoch": 2.807142857142857,
2773
+ "grad_norm": 3.37313175201416,
2774
+ "learning_rate": 6.428571428571429e-06,
2775
+ "loss": 0.3156,
2776
+ "step": 3930
2777
+ },
2778
+ {
2779
+ "epoch": 2.814285714285714,
2780
+ "grad_norm": 4.180731773376465,
2781
+ "learning_rate": 6.190476190476191e-06,
2782
+ "loss": 0.279,
2783
+ "step": 3940
2784
+ },
2785
+ {
2786
+ "epoch": 2.821428571428571,
2787
+ "grad_norm": 5.761970520019531,
2788
+ "learning_rate": 5.9523809523809525e-06,
2789
+ "loss": 0.2478,
2790
+ "step": 3950
2791
+ },
2792
+ {
2793
+ "epoch": 2.8285714285714287,
2794
+ "grad_norm": 3.018902063369751,
2795
+ "learning_rate": 5.7142857142857145e-06,
2796
+ "loss": 0.2224,
2797
+ "step": 3960
2798
+ },
2799
+ {
2800
+ "epoch": 2.835714285714286,
2801
+ "grad_norm": 2.3872735500335693,
2802
+ "learning_rate": 5.4761904761904765e-06,
2803
+ "loss": 0.1913,
2804
+ "step": 3970
2805
+ },
2806
+ {
2807
+ "epoch": 2.842857142857143,
2808
+ "grad_norm": 4.139348983764648,
2809
+ "learning_rate": 5.2380952380952384e-06,
2810
+ "loss": 0.2815,
2811
+ "step": 3980
2812
+ },
2813
+ {
2814
+ "epoch": 2.85,
2815
+ "grad_norm": 7.203916072845459,
2816
+ "learning_rate": 5e-06,
2817
+ "loss": 0.3189,
2818
+ "step": 3990
2819
+ },
2820
+ {
2821
+ "epoch": 2.857142857142857,
2822
+ "grad_norm": 5.5615715980529785,
2823
+ "learning_rate": 4.7619047619047615e-06,
2824
+ "loss": 0.2584,
2825
+ "step": 4000
2826
+ },
2827
+ {
2828
+ "epoch": 2.8642857142857143,
2829
+ "grad_norm": 1.6023494005203247,
2830
+ "learning_rate": 4.5238095238095235e-06,
2831
+ "loss": 0.2672,
2832
+ "step": 4010
2833
+ },
2834
+ {
2835
+ "epoch": 2.8714285714285714,
2836
+ "grad_norm": 5.98526668548584,
2837
+ "learning_rate": 4.285714285714286e-06,
2838
+ "loss": 0.2492,
2839
+ "step": 4020
2840
+ },
2841
+ {
2842
+ "epoch": 2.8785714285714286,
2843
+ "grad_norm": 2.8703975677490234,
2844
+ "learning_rate": 4.047619047619048e-06,
2845
+ "loss": 0.2164,
2846
+ "step": 4030
2847
+ },
2848
+ {
2849
+ "epoch": 2.8857142857142857,
2850
+ "grad_norm": 4.569535732269287,
2851
+ "learning_rate": 3.8095238095238102e-06,
2852
+ "loss": 0.2991,
2853
+ "step": 4040
2854
+ },
2855
+ {
2856
+ "epoch": 2.892857142857143,
2857
+ "grad_norm": 3.0604588985443115,
2858
+ "learning_rate": 3.5714285714285714e-06,
2859
+ "loss": 0.2416,
2860
+ "step": 4050
2861
+ },
2862
+ {
2863
+ "epoch": 2.9,
2864
+ "grad_norm": 3.326860189437866,
2865
+ "learning_rate": 3.3333333333333333e-06,
2866
+ "loss": 0.2035,
2867
+ "step": 4060
2868
+ },
2869
+ {
2870
+ "epoch": 2.907142857142857,
2871
+ "grad_norm": 4.7924675941467285,
2872
+ "learning_rate": 3.0952380952380953e-06,
2873
+ "loss": 0.3382,
2874
+ "step": 4070
2875
+ },
2876
+ {
2877
+ "epoch": 2.914285714285714,
2878
+ "grad_norm": 1.4087116718292236,
2879
+ "learning_rate": 2.8571428571428573e-06,
2880
+ "loss": 0.1943,
2881
+ "step": 4080
2882
+ },
2883
+ {
2884
+ "epoch": 2.9214285714285713,
2885
+ "grad_norm": 2.154977798461914,
2886
+ "learning_rate": 2.6190476190476192e-06,
2887
+ "loss": 0.2088,
2888
+ "step": 4090
2889
+ },
2890
+ {
2891
+ "epoch": 2.928571428571429,
2892
+ "grad_norm": 2.5638303756713867,
2893
+ "learning_rate": 2.3809523809523808e-06,
2894
+ "loss": 0.2024,
2895
+ "step": 4100
2896
+ },
2897
+ {
2898
+ "epoch": 2.935714285714286,
2899
+ "grad_norm": 5.622275352478027,
2900
+ "learning_rate": 2.142857142857143e-06,
2901
+ "loss": 0.2493,
2902
+ "step": 4110
2903
+ },
2904
+ {
2905
+ "epoch": 2.942857142857143,
2906
+ "grad_norm": 4.387233734130859,
2907
+ "learning_rate": 1.9047619047619051e-06,
2908
+ "loss": 0.2102,
2909
+ "step": 4120
2910
+ },
2911
+ {
2912
+ "epoch": 2.95,
2913
+ "grad_norm": 3.105154514312744,
2914
+ "learning_rate": 1.6666666666666667e-06,
2915
+ "loss": 0.2266,
2916
+ "step": 4130
2917
+ },
2918
+ {
2919
+ "epoch": 2.9571428571428573,
2920
+ "grad_norm": 2.8728506565093994,
2921
+ "learning_rate": 1.4285714285714286e-06,
2922
+ "loss": 0.1978,
2923
+ "step": 4140
2924
+ },
2925
+ {
2926
+ "epoch": 2.9642857142857144,
2927
+ "grad_norm": 4.60084867477417,
2928
+ "learning_rate": 1.1904761904761904e-06,
2929
+ "loss": 0.2214,
2930
+ "step": 4150
2931
+ },
2932
+ {
2933
+ "epoch": 2.9714285714285715,
2934
+ "grad_norm": 3.7164621353149414,
2935
+ "learning_rate": 9.523809523809526e-07,
2936
+ "loss": 0.2272,
2937
+ "step": 4160
2938
+ },
2939
+ {
2940
+ "epoch": 2.9785714285714286,
2941
+ "grad_norm": 2.304844856262207,
2942
+ "learning_rate": 7.142857142857143e-07,
2943
+ "loss": 0.2652,
2944
+ "step": 4170
2945
+ },
2946
+ {
2947
+ "epoch": 2.9857142857142858,
2948
+ "grad_norm": 2.9933526515960693,
2949
+ "learning_rate": 4.761904761904763e-07,
2950
+ "loss": 0.2606,
2951
+ "step": 4180
2952
+ },
2953
+ {
2954
+ "epoch": 2.992857142857143,
2955
+ "grad_norm": 2.8444643020629883,
2956
+ "learning_rate": 2.3809523809523814e-07,
2957
+ "loss": 0.2501,
2958
+ "step": 4190
2959
+ },
2960
+ {
2961
+ "epoch": 3.0,
2962
+ "grad_norm": 3.5711305141448975,
2963
+ "learning_rate": 0.0,
2964
+ "loss": 0.2887,
2965
+ "step": 4200
2966
+ },
2967
+ {
2968
+ "epoch": 3.0,
2969
+ "eval_loss": 0.23848384618759155,
2970
+ "eval_runtime": 68.3028,
2971
+ "eval_samples_per_second": 2.928,
2972
+ "eval_steps_per_second": 0.366,
2973
+ "step": 4200
2974
+ }
2975
+ ],
2976
+ "logging_steps": 10,
2977
+ "max_steps": 4200,
2978
+ "num_input_tokens_seen": 0,
2979
+ "num_train_epochs": 3,
2980
+ "save_steps": 500,
2981
+ "stateful_callbacks": {
2982
+ "TrainerControl": {
2983
+ "args": {
2984
+ "should_epoch_stop": false,
2985
+ "should_evaluate": false,
2986
+ "should_log": false,
2987
+ "should_save": true,
2988
+ "should_training_stop": true
2989
+ },
2990
+ "attributes": {}
2991
+ }
2992
+ },
2993
+ "total_flos": 2.5133825654784e+16,
2994
+ "train_batch_size": 2,
2995
+ "trial_name": null,
2996
+ "trial_params": null
2997
+ }
checkpoint-4200/training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aef410e0e1d0110bb1523d95ae4f80f726beb1aea9164ada6672ae6ffeddc06b
3
+ size 5240
tokenizer.json CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:fbe07b35a61be0ec850f12e9379be8ca47b4ef845b20f0d146a8a2bfdb0e7e3e
3
  size 17210195
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c70650b4236027dc8db4abca6b918783a8ed2ee38cd69142f6dbbeb5945f876f
3
  size 17210195
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1982b07d8277b0bc7aaf4f34307bfaa24bd97b8994cd2c48e972769e68324f20
3
  size 5240
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:aef410e0e1d0110bb1523d95ae4f80f726beb1aea9164ada6672ae6ffeddc06b
3
  size 5240