Hugging Face's logo

llama-duo
/
gemma7b-summarize-gpt4o-8k

PEFT
TensorBoard
Safetensors
gemma
alignment-handbook
Generated from Trainer
4-bit precision
bitsandbytes
Model card Files Metrics Community
juyongjiang commited on
Commit
dcd490a
•
1 Parent(s): e451697

update model checkpoint

Browse files
Files changed (11) hide show
  1. README.md +17 -20
  2. adapter_config.json +0 -5
  3. adapter_model.safetensors +2 -2
  4. all_results.json +10 -10
  5. config.json +2 -2
  6. eval_results.json +4 -4
  7. runs/Jun13_05-51-46_gpu1-2/events.out.tfevents.1718229170.gpu1-2.1122759.0 +3 -0
  8. runs/Jun13_05-51-46_gpu1-2/events.out.tfevents.1718229510.gpu1-2.1122759.1 +3 -0
  9. train_results.json +6 -6
  10. trainer_state.json +197 -715
  11. training_args.bin +1 -1
README.md CHANGED
@@ -1,14 +1,11 @@
1
  ---
2
- license: gemma
3
  library_name: peft
4
  tags:
5
  - alignment-handbook
6
- - trl
7
- - sft
8
  - generated_from_trainer
9
- base_model: google/gemma-7b
10
  datasets:
11
  - llama-duo/synth_summarize_dataset_dedup
 
12
  model-index:
13
  - name: gemma7b-summarize-gpt4o-8k
14
  results: []
@@ -21,7 +18,7 @@ should probably proofread and complete it, then remove this comment. -->
21
 
22
  This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the llama-duo/synth_summarize_dataset_dedup dataset.
23
  It achieves the following results on the evaluation set:
24
- - Loss: 4.5136
25
 
26
  ## Model description
27
 
@@ -42,13 +39,13 @@ More information needed
42
  The following hyperparameters were used during training:
43
  - learning_rate: 0.0002
44
  - train_batch_size: 4
45
- - eval_batch_size: 4
46
  - seed: 42
47
  - distributed_type: multi-GPU
48
- - num_devices: 2
49
  - gradient_accumulation_steps: 2
50
- - total_train_batch_size: 16
51
- - total_eval_batch_size: 8
52
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
53
  - lr_scheduler_type: cosine
54
  - lr_scheduler_warmup_ratio: 0.1
@@ -58,22 +55,22 @@ The following hyperparameters were used during training:
58
 
59
  | Training Loss | Epoch | Step | Validation Loss |
60
  |:-------------:|:-----:|:----:|:---------------:|
61
- | 1.7584 | 1.0 | 51 | 2.7736 |
62
- | 1.0801 | 2.0 | 102 | 2.5391 |
63
- | 0.9341 | 3.0 | 153 | 2.5877 |
64
- | 0.7796 | 4.0 | 204 | 2.6409 |
65
- | 0.6583 | 5.0 | 255 | 2.8397 |
66
- | 0.4945 | 6.0 | 306 | 3.1891 |
67
- | 0.3891 | 7.0 | 357 | 3.7245 |
68
- | 0.3029 | 8.0 | 408 | 4.1951 |
69
- | 0.259 | 9.0 | 459 | 4.4996 |
70
- | 0.2532 | 10.0 | 510 | 4.5136 |
71
 
72
 
73
  ### Framework versions
74
 
75
  - PEFT 0.10.0
76
  - Transformers 4.40.0
77
- - Pytorch 2.2.1+cu121
78
  - Datasets 2.18.0
79
  - Tokenizers 0.19.1
 
1
  ---
 
2
  library_name: peft
3
  tags:
4
  - alignment-handbook
 
 
5
  - generated_from_trainer
 
6
  datasets:
7
  - llama-duo/synth_summarize_dataset_dedup
8
+ base_model: google/gemma-7b
9
  model-index:
10
  - name: gemma7b-summarize-gpt4o-8k
11
  results: []
 
18
 
19
  This model is a fine-tuned version of [google/gemma-7b](https://huggingface.co/google/gemma-7b) on the llama-duo/synth_summarize_dataset_dedup dataset.
20
  It achieves the following results on the evaluation set:
21
+ - Loss: 2.8129
22
 
23
  ## Model description
24
 
 
39
  The following hyperparameters were used during training:
40
  - learning_rate: 0.0002
41
  - train_batch_size: 4
42
+ - eval_batch_size: 2
43
  - seed: 42
44
  - distributed_type: multi-GPU
45
+ - num_devices: 8
46
  - gradient_accumulation_steps: 2
47
+ - total_train_batch_size: 64
48
+ - total_eval_batch_size: 16
49
  - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
50
  - lr_scheduler_type: cosine
51
  - lr_scheduler_warmup_ratio: 0.1
 
55
 
56
  | Training Loss | Epoch | Step | Validation Loss |
57
  |:-------------:|:-----:|:----:|:---------------:|
58
+ | 30.8653 | 1.0 | 14 | 10.6638 |
59
+ | 18.5328 | 2.0 | 28 | 7.3031 |
60
+ | 11.486 | 3.0 | 42 | 6.6280 |
61
+ | 2.4959 | 4.0 | 56 | 3.5087 |
62
+ | 1.742 | 5.0 | 70 | 3.0216 |
63
+ | 1.5971 | 6.0 | 84 | 2.8802 |
64
+ | 1.4792 | 7.0 | 98 | 2.8307 |
65
+ | 1.4333 | 8.0 | 112 | 2.8081 |
66
+ | 1.4129 | 9.0 | 126 | 2.8151 |
67
+ | 1.4048 | 10.0 | 140 | 2.8129 |
68
 
69
 
70
  ### Framework versions
71
 
72
  - PEFT 0.10.0
73
  - Transformers 4.40.0
74
+ - Pytorch 2.1.2+cu121
75
  - Datasets 2.18.0
76
  - Tokenizers 0.19.1
adapter_config.json CHANGED
@@ -20,12 +20,7 @@
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
23
- "k_proj",
24
- "o_proj",
25
  "q_proj",
26
- "up_proj",
27
- "down_proj",
28
- "gate_proj",
29
  "v_proj"
30
  ],
31
  "task_type": "CAUSAL_LM",
 
20
  "rank_pattern": {},
21
  "revision": null,
22
  "target_modules": [
 
 
23
  "q_proj",
 
 
 
24
  "v_proj"
25
  ],
26
  "task_type": "CAUSAL_LM",
adapter_model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:e37c72c8e377066f1fd2d65bc7e6357ffd561b096563a48fb7260a9d34b9fd33
3
- size 50056096
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8bc2f19402ce8afd67aef1677e88a99231111e7c9580b7fa9292902e001d5df5
3
+ size 6437384
all_results.json CHANGED
@@ -1,14 +1,14 @@
1
  {
2
  "epoch": 10.0,
3
- "eval_loss": 4.513559818267822,
4
- "eval_runtime": 2.0633,
5
  "eval_samples": 25,
6
- "eval_samples_per_second": 4.847,
7
- "eval_steps_per_second": 0.969,
8
- "total_flos": 3.8986265891805594e+17,
9
- "train_loss": 2.3579357918571024,
10
- "train_runtime": 3338.7641,
11
- "train_samples": 7499,
12
- "train_samples_per_second": 2.441,
13
- "train_steps_per_second": 0.153
14
  }
 
1
  {
2
  "epoch": 10.0,
3
+ "eval_loss": 2.812880277633667,
4
+ "eval_runtime": 0.2342,
5
  "eval_samples": 25,
6
+ "eval_samples_per_second": 42.691,
7
+ "eval_steps_per_second": 4.269,
8
+ "total_flos": 4.268849030789857e+17,
9
+ "train_loss": 7.876109651156834,
10
+ "train_runtime": 340.0833,
11
+ "train_samples": 8076,
12
+ "train_samples_per_second": 25.758,
13
+ "train_steps_per_second": 0.412
14
  }
config.json CHANGED
@@ -23,9 +23,9 @@
23
  "_load_in_4bit": true,
24
  "_load_in_8bit": false,
25
  "bnb_4bit_compute_dtype": "bfloat16",
26
- "bnb_4bit_quant_storage": "bfloat16",
27
  "bnb_4bit_quant_type": "nf4",
28
- "bnb_4bit_use_double_quant": true,
29
  "llm_int8_enable_fp32_cpu_offload": false,
30
  "llm_int8_has_fp16_weight": false,
31
  "llm_int8_skip_modules": null,
 
23
  "_load_in_4bit": true,
24
  "_load_in_8bit": false,
25
  "bnb_4bit_compute_dtype": "bfloat16",
26
+ "bnb_4bit_quant_storage": "uint8",
27
  "bnb_4bit_quant_type": "nf4",
28
+ "bnb_4bit_use_double_quant": false,
29
  "llm_int8_enable_fp32_cpu_offload": false,
30
  "llm_int8_has_fp16_weight": false,
31
  "llm_int8_skip_modules": null,
eval_results.json CHANGED
@@ -1,8 +1,8 @@
1
  {
2
  "epoch": 10.0,
3
- "eval_loss": 4.513559818267822,
4
- "eval_runtime": 2.0633,
5
  "eval_samples": 25,
6
- "eval_samples_per_second": 4.847,
7
- "eval_steps_per_second": 0.969
8
  }
 
1
  {
2
  "epoch": 10.0,
3
+ "eval_loss": 2.812880277633667,
4
+ "eval_runtime": 0.2342,
5
  "eval_samples": 25,
6
+ "eval_samples_per_second": 42.691,
7
+ "eval_steps_per_second": 4.269
8
  }
runs/Jun13_05-51-46_gpu1-2/events.out.tfevents.1718229170.gpu1-2.1122759.0 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c8e1848e9983bc1ed6b9bf0c5970b363dca4d2437ff9836fe4c5cf616d9581a7
3
+ size 14481
runs/Jun13_05-51-46_gpu1-2/events.out.tfevents.1718229510.gpu1-2.1122759.1 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a26bcebc6fd300f9589badc66fa7a3a6f5baba4444639083285fc3c03dd98122
3
+ size 359
train_results.json CHANGED
@@ -1,9 +1,9 @@
1
  {
2
  "epoch": 10.0,
3
- "total_flos": 3.8986265891805594e+17,
4
- "train_loss": 2.3579357918571024,
5
- "train_runtime": 3338.7641,
6
- "train_samples": 7499,
7
- "train_samples_per_second": 2.441,
8
- "train_steps_per_second": 0.153
9
  }
 
1
  {
2
  "epoch": 10.0,
3
+ "total_flos": 4.268849030789857e+17,
4
+ "train_loss": 7.876109651156834,
5
+ "train_runtime": 340.0833,
6
+ "train_samples": 8076,
7
+ "train_samples_per_second": 25.758,
8
+ "train_steps_per_second": 0.412
9
  }
trainer_state.json CHANGED
@@ -3,828 +3,310 @@
3
  "best_model_checkpoint": null,
4
  "epoch": 10.0,
5
  "eval_steps": 500,
6
- "global_step": 510,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
- "epoch": 0.0196078431372549,
13
- "grad_norm": 494.0,
14
- "learning_rate": 3.92156862745098e-06,
15
- "loss": 48.0011,
16
  "step": 1
17
  },
18
  {
19
- "epoch": 0.09803921568627451,
20
- "grad_norm": 394.0,
21
- "learning_rate": 1.9607843137254903e-05,
22
- "loss": 46.237,
23
  "step": 5
24
  },
25
  {
26
- "epoch": 0.19607843137254902,
27
- "grad_norm": 131.0,
28
- "learning_rate": 3.9215686274509805e-05,
29
- "loss": 36.1065,
30
  "step": 10
31
  },
32
- {
33
- "epoch": 0.29411764705882354,
34
- "grad_norm": 36.25,
35
- "learning_rate": 5.882352941176471e-05,
36
- "loss": 23.9218,
37
- "step": 15
38
- },
39
- {
40
- "epoch": 0.39215686274509803,
41
- "grad_norm": 11.6875,
42
- "learning_rate": 7.843137254901961e-05,
43
- "loss": 20.3348,
44
- "step": 20
45
- },
46
- {
47
- "epoch": 0.49019607843137253,
48
- "grad_norm": 8.125,
49
- "learning_rate": 9.80392156862745e-05,
50
- "loss": 19.5012,
51
- "step": 25
52
- },
53
- {
54
- "epoch": 0.5882352941176471,
55
- "grad_norm": 20.375,
56
- "learning_rate": 0.00011764705882352942,
57
- "loss": 17.6271,
58
- "step": 30
59
- },
60
- {
61
- "epoch": 0.6862745098039216,
62
- "grad_norm": 60.0,
63
- "learning_rate": 0.0001372549019607843,
64
- "loss": 13.6754,
65
- "step": 35
66
- },
67
- {
68
- "epoch": 0.7843137254901961,
69
- "grad_norm": 11.0625,
70
- "learning_rate": 0.00015686274509803922,
71
- "loss": 4.1094,
72
- "step": 40
73
- },
74
- {
75
- "epoch": 0.8823529411764706,
76
- "grad_norm": 5.0,
77
- "learning_rate": 0.00017647058823529413,
78
- "loss": 2.136,
79
- "step": 45
80
- },
81
- {
82
- "epoch": 0.9803921568627451,
83
- "grad_norm": 3.375,
84
- "learning_rate": 0.000196078431372549,
85
- "loss": 1.7584,
86
- "step": 50
87
- },
88
  {
89
  "epoch": 1.0,
90
- "eval_loss": 2.7735681533813477,
91
- "eval_runtime": 2.0507,
92
- "eval_samples_per_second": 4.876,
93
- "eval_steps_per_second": 0.975,
94
- "step": 51
95
  },
96
  {
97
- "epoch": 1.0784313725490196,
98
- "grad_norm": 2.6875,
99
- "learning_rate": 0.00019996252538028507,
100
- "loss": 1.5368,
101
- "step": 55
102
- },
103
- {
104
- "epoch": 1.1764705882352942,
105
- "grad_norm": 23.125,
106
- "learning_rate": 0.00019981033287370443,
107
- "loss": 1.4659,
108
- "step": 60
109
- },
110
- {
111
- "epoch": 1.2745098039215685,
112
- "grad_norm": 3.640625,
113
- "learning_rate": 0.00019954125840299165,
114
- "loss": 1.3602,
115
- "step": 65
116
- },
117
- {
118
- "epoch": 1.3725490196078431,
119
- "grad_norm": 6.40625,
120
- "learning_rate": 0.00019915561706530883,
121
- "loss": 1.3083,
122
- "step": 70
123
- },
124
- {
125
- "epoch": 1.4705882352941178,
126
- "grad_norm": 2.53125,
127
- "learning_rate": 0.00019865386046236596,
128
- "loss": 1.2742,
129
- "step": 75
130
- },
131
- {
132
- "epoch": 1.5686274509803921,
133
- "grad_norm": 3.21875,
134
- "learning_rate": 0.0001980365761715769,
135
- "loss": 1.2384,
136
- "step": 80
137
- },
138
- {
139
- "epoch": 1.6666666666666665,
140
- "grad_norm": 3.8125,
141
- "learning_rate": 0.00019730448705798239,
142
- "loss": 1.1832,
143
- "step": 85
144
- },
145
- {
146
- "epoch": 1.7647058823529411,
147
- "grad_norm": 2.953125,
148
- "learning_rate": 0.00019645845042774553,
149
- "loss": 1.1402,
150
- "step": 90
151
  },
152
  {
153
- "epoch": 1.8627450980392157,
154
- "grad_norm": 4.6875,
155
- "learning_rate": 0.00019549945702421144,
156
- "loss": 1.0986,
157
- "step": 95
158
  },
159
  {
160
- "epoch": 1.9607843137254903,
161
- "grad_norm": 2.015625,
162
- "learning_rate": 0.00019442862986770646,
163
- "loss": 1.0801,
164
- "step": 100
165
  },
166
  {
167
  "epoch": 2.0,
168
- "eval_loss": 2.539125442504883,
169
- "eval_runtime": 2.0419,
170
- "eval_samples_per_second": 4.897,
171
- "eval_steps_per_second": 0.979,
172
- "step": 102
173
  },
174
  {
175
- "epoch": 2.0588235294117645,
176
- "grad_norm": 42.0,
177
- "learning_rate": 0.00019324722294043558,
178
- "loss": 1.0191,
179
- "step": 105
180
- },
181
- {
182
- "epoch": 2.156862745098039,
183
- "grad_norm": 3.546875,
184
- "learning_rate": 0.00019195661971801827,
185
- "loss": 0.9673,
186
- "step": 110
187
- },
188
- {
189
- "epoch": 2.2549019607843137,
190
- "grad_norm": 1.1484375,
191
- "learning_rate": 0.00019055833154938207,
192
- "loss": 0.9712,
193
- "step": 115
194
- },
195
- {
196
- "epoch": 2.3529411764705883,
197
- "grad_norm": 1.671875,
198
- "learning_rate": 0.00018905399588691163,
199
- "loss": 0.9414,
200
- "step": 120
201
- },
202
- {
203
- "epoch": 2.450980392156863,
204
- "grad_norm": 1.46875,
205
- "learning_rate": 0.00018744537436892516,
206
- "loss": 0.9309,
207
- "step": 125
208
- },
209
- {
210
- "epoch": 2.549019607843137,
211
- "grad_norm": 2.359375,
212
- "learning_rate": 0.00018573435075672424,
213
- "loss": 0.9278,
214
- "step": 130
215
- },
216
- {
217
- "epoch": 2.6470588235294117,
218
- "grad_norm": 1.5625,
219
- "learning_rate": 0.00018392292872863267,
220
- "loss": 0.928,
221
- "step": 135
222
- },
223
- {
224
- "epoch": 2.7450980392156863,
225
- "grad_norm": 4.6875,
226
- "learning_rate": 0.00018201322953360758,
227
- "loss": 0.9431,
228
- "step": 140
229
  },
230
  {
231
- "epoch": 2.843137254901961,
232
- "grad_norm": 1.7578125,
233
- "learning_rate": 0.00018000748950717038,
234
- "loss": 0.934,
235
- "step": 145
236
  },
237
  {
238
- "epoch": 2.9411764705882355,
239
- "grad_norm": 4.8125,
240
- "learning_rate": 0.00017790805745256704,
241
- "loss": 0.9341,
242
- "step": 150
243
  },
244
  {
245
  "epoch": 3.0,
246
- "eval_loss": 2.587712049484253,
247
- "eval_runtime": 2.0445,
248
- "eval_samples_per_second": 4.891,
249
- "eval_steps_per_second": 0.978,
250
- "step": 153
251
  },
252
  {
253
- "epoch": 3.0392156862745097,
254
- "grad_norm": 1.2578125,
255
- "learning_rate": 0.00017571739189022365,
256
- "loss": 0.8748,
257
- "step": 155
258
- },
259
- {
260
- "epoch": 3.1372549019607843,
261
- "grad_norm": 1.765625,
262
- "learning_rate": 0.00017343805817871886,
263
- "loss": 0.805,
264
- "step": 160
265
- },
266
- {
267
- "epoch": 3.235294117647059,
268
- "grad_norm": 1.828125,
269
- "learning_rate": 0.00017107272551064473,
270
- "loss": 0.7843,
271
- "step": 165
272
- },
273
- {
274
- "epoch": 3.3333333333333335,
275
- "grad_norm": 5.40625,
276
- "learning_rate": 0.0001686241637868734,
277
- "loss": 0.7917,
278
- "step": 170
279
- },
280
- {
281
- "epoch": 3.431372549019608,
282
- "grad_norm": 1.21875,
283
- "learning_rate": 0.00016609524037289019,
284
- "loss": 0.8003,
285
- "step": 175
286
- },
287
- {
288
- "epoch": 3.5294117647058822,
289
- "grad_norm": 0.94140625,
290
- "learning_rate": 0.0001634889167409923,
291
- "loss": 0.7639,
292
- "step": 180
293
- },
294
- {
295
- "epoch": 3.627450980392157,
296
- "grad_norm": 1.3046875,
297
- "learning_rate": 0.00016080824500228367,
298
- "loss": 0.7796,
299
- "step": 185
300
- },
301
- {
302
- "epoch": 3.7254901960784315,
303
- "grad_norm": 1.8515625,
304
- "learning_rate": 0.00015805636433252891,
305
- "loss": 0.7783,
306
- "step": 190
307
  },
308
  {
309
- "epoch": 3.8235294117647056,
310
- "grad_norm": 1.265625,
311
- "learning_rate": 0.0001552364972960506,
312
- "loss": 0.7854,
313
- "step": 195
314
  },
315
  {
316
- "epoch": 3.9215686274509802,
317
- "grad_norm": 2.9375,
318
- "learning_rate": 0.00015235194607197508,
319
- "loss": 0.7796,
320
- "step": 200
321
  },
322
  {
323
  "epoch": 4.0,
324
- "eval_loss": 2.6408610343933105,
325
- "eval_runtime": 2.0419,
326
- "eval_samples_per_second": 4.897,
327
- "eval_steps_per_second": 0.979,
328
- "step": 204
329
- },
330
- {
331
- "epoch": 4.019607843137255,
332
- "grad_norm": 1.8828125,
333
- "learning_rate": 0.0001494060885872464,
334
- "loss": 0.7823,
335
- "step": 205
336
- },
337
- {
338
- "epoch": 4.117647058823529,
339
- "grad_norm": 16.875,
340
- "learning_rate": 0.00014640237456093634,
341
- "loss": 0.6465,
342
- "step": 210
343
- },
344
- {
345
- "epoch": 4.215686274509804,
346
- "grad_norm": 1.0625,
347
- "learning_rate": 0.00014334432146448272,
348
- "loss": 0.621,
349
- "step": 215
350
- },
351
- {
352
- "epoch": 4.313725490196078,
353
- "grad_norm": 1.5625,
354
- "learning_rate": 0.00014023551040258725,
355
- "loss": 0.612,
356
- "step": 220
357
- },
358
- {
359
- "epoch": 4.411764705882353,
360
- "grad_norm": 1.3828125,
361
- "learning_rate": 0.00013707958191959608,
362
- "loss": 0.6456,
363
- "step": 225
364
- },
365
- {
366
- "epoch": 4.509803921568627,
367
- "grad_norm": 1.375,
368
- "learning_rate": 0.00013388023173627414,
369
- "loss": 0.6361,
370
- "step": 230
371
- },
372
- {
373
- "epoch": 4.607843137254902,
374
- "grad_norm": 1.0234375,
375
- "learning_rate": 0.00013064120642196548,
376
- "loss": 0.6487,
377
- "step": 235
378
- },
379
- {
380
- "epoch": 4.705882352941177,
381
- "grad_norm": 0.921875,
382
- "learning_rate": 0.0001273662990072083,
383
- "loss": 0.6424,
384
- "step": 240
385
  },
386
  {
387
- "epoch": 4.803921568627451,
388
- "grad_norm": 0.8984375,
389
- "learning_rate": 0.00012405934454194146,
390
- "loss": 0.653,
391
- "step": 245
392
  },
393
  {
394
- "epoch": 4.901960784313726,
395
- "grad_norm": 0.953125,
396
- "learning_rate": 0.00012072421560450497,
397
- "loss": 0.6493,
398
- "step": 250
399
  },
400
  {
401
  "epoch": 5.0,
402
- "grad_norm": 1.0703125,
403
  "learning_rate": 0.00011736481776669306,
404
- "loss": 0.6583,
405
- "step": 255
406
  },
407
  {
408
  "epoch": 5.0,
409
- "eval_loss": 2.8397467136383057,
410
- "eval_runtime": 2.0446,
411
- "eval_samples_per_second": 4.891,
412
- "eval_steps_per_second": 0.978,
413
- "step": 255
414
- },
415
- {
416
- "epoch": 5.098039215686274,
417
- "grad_norm": 1.9609375,
418
- "learning_rate": 0.00011398508502017047,
419
- "loss": 0.5107,
420
- "step": 260
421
- },
422
- {
423
- "epoch": 5.196078431372549,
424
- "grad_norm": 1.09375,
425
- "learning_rate": 0.00011058897516960816,
426
- "loss": 0.4939,
427
- "step": 265
428
- },
429
- {
430
- "epoch": 5.294117647058823,
431
- "grad_norm": 1.0859375,
432
- "learning_rate": 0.00010718046519793276,
433
- "loss": 0.4931,
434
- "step": 270
435
- },
436
- {
437
- "epoch": 5.392156862745098,
438
- "grad_norm": 1.1328125,
439
- "learning_rate": 0.00010376354660911771,
440
- "loss": 0.4983,
441
- "step": 275
442
- },
443
- {
444
- "epoch": 5.490196078431373,
445
- "grad_norm": 1.0390625,
446
- "learning_rate": 0.00010034222075396953,
447
- "loss": 0.5112,
448
- "step": 280
449
- },
450
- {
451
- "epoch": 5.588235294117647,
452
- "grad_norm": 1.0546875,
453
- "learning_rate": 9.692049414438299e-05,
454
- "loss": 0.4841,
455
- "step": 285
456
- },
457
- {
458
- "epoch": 5.686274509803922,
459
- "grad_norm": 1.0078125,
460
- "learning_rate": 9.350237376155269e-05,
461
- "loss": 0.4976,
462
- "step": 290
463
- },
464
- {
465
- "epoch": 5.784313725490196,
466
- "grad_norm": 1.0390625,
467
- "learning_rate": 9.00918623636349e-05,
468
- "loss": 0.4954,
469
- "step": 295
470
  },
471
  {
472
- "epoch": 5.882352941176471,
473
- "grad_norm": 1.125,
474
- "learning_rate": 8.669295379835467e-05,
475
- "loss": 0.4989,
476
- "step": 300
477
  },
478
  {
479
- "epoch": 5.980392156862745,
480
- "grad_norm": 0.96484375,
481
- "learning_rate": 8.330962832604747e-05,
482
- "loss": 0.4945,
483
- "step": 305
484
  },
485
  {
486
  "epoch": 6.0,
487
- "eval_loss": 3.189112424850464,
488
- "eval_runtime": 2.0474,
489
- "eval_samples_per_second": 4.884,
490
- "eval_steps_per_second": 0.977,
491
- "step": 306
492
- },
493
- {
494
- "epoch": 6.078431372549019,
495
- "grad_norm": 1.03125,
496
- "learning_rate": 7.994584795861249e-05,
497
- "loss": 0.4204,
498
- "step": 310
499
- },
500
- {
501
- "epoch": 6.176470588235294,
502
- "grad_norm": 1.109375,
503
- "learning_rate": 7.660555181983518e-05,
504
- "loss": 0.3844,
505
- "step": 315
506
- },
507
- {
508
- "epoch": 6.2745098039215685,
509
- "grad_norm": 1.21875,
510
- "learning_rate": 7.329265153251285e-05,
511
- "loss": 0.3839,
512
- "step": 320
513
- },
514
- {
515
- "epoch": 6.372549019607844,
516
- "grad_norm": 0.9453125,
517
- "learning_rate": 7.001102663778533e-05,
518
- "loss": 0.3772,
519
- "step": 325
520
  },
521
  {
522
- "epoch": 6.470588235294118,
523
- "grad_norm": 0.99609375,
524
- "learning_rate": 6.676452005203406e-05,
525
- "loss": 0.3877,
526
- "step": 330
527
- },
528
- {
529
- "epoch": 6.568627450980392,
530
- "grad_norm": 1.0078125,
531
- "learning_rate": 6.355693356667065e-05,
532
- "loss": 0.388,
533
- "step": 335
534
- },
535
- {
536
- "epoch": 6.666666666666667,
537
- "grad_norm": 0.98046875,
538
- "learning_rate": 6.039202339608432e-05,
539
- "loss": 0.3832,
540
- "step": 340
541
- },
542
- {
543
- "epoch": 6.764705882352941,
544
- "grad_norm": 1.0859375,
545
- "learning_rate": 5.727349577896194e-05,
546
- "loss": 0.3926,
547
- "step": 345
548
  },
549
  {
550
- "epoch": 6.862745098039216,
551
- "grad_norm": 1.046875,
552
- "learning_rate": 5.4205002638131406e-05,
553
- "loss": 0.387,
554
- "step": 350
555
  },
556
  {
557
- "epoch": 6.96078431372549,
558
- "grad_norm": 0.98046875,
559
- "learning_rate": 5.119013730401152e-05,
560
- "loss": 0.3891,
561
- "step": 355
562
  },
563
  {
564
  "epoch": 7.0,
565
- "eval_loss": 3.7245430946350098,
566
- "eval_runtime": 2.0488,
567
- "eval_samples_per_second": 4.881,
568
- "eval_steps_per_second": 0.976,
569
- "step": 357
570
- },
571
- {
572
- "epoch": 7.0588235294117645,
573
- "grad_norm": 0.81640625,
574
- "learning_rate": 4.823243030667576e-05,
575
- "loss": 0.3491,
576
- "step": 360
577
- },
578
- {
579
- "epoch": 7.1568627450980395,
580
- "grad_norm": 1.0546875,
581
- "learning_rate": 4.533534524145756e-05,
582
- "loss": 0.3038,
583
- "step": 365
584
- },
585
- {
586
- "epoch": 7.254901960784314,
587
- "grad_norm": 0.91015625,
588
- "learning_rate": 4.250227471293935e-05,
589
- "loss": 0.3093,
590
- "step": 370
591
- },
592
- {
593
- "epoch": 7.352941176470588,
594
- "grad_norm": 0.99609375,
595
- "learning_rate": 3.973653636207437e-05,
596
- "loss": 0.3096,
597
- "step": 375
598
- },
599
- {
600
- "epoch": 7.450980392156863,
601
- "grad_norm": 0.94921875,
602
- "learning_rate": 3.7041368981094024e-05,
603
- "loss": 0.3137,
604
- "step": 380
605
  },
606
  {
607
- "epoch": 7.549019607843137,
608
- "grad_norm": 0.984375,
609
- "learning_rate": 3.441992872075027e-05,
610
- "loss": 0.3073,
611
- "step": 385
612
- },
613
- {
614
- "epoch": 7.647058823529412,
615
- "grad_norm": 0.9375,
616
- "learning_rate": 3.187528539433458e-05,
617
- "loss": 0.2992,
618
- "step": 390
619
- },
620
- {
621
- "epoch": 7.745098039215686,
622
- "grad_norm": 0.93359375,
623
- "learning_rate": 2.9410418882801683e-05,
624
- "loss": 0.3063,
625
- "step": 395
626
  },
627
  {
628
- "epoch": 7.8431372549019605,
629
- "grad_norm": 0.94140625,
630
- "learning_rate": 2.702821564520732e-05,
631
- "loss": 0.308,
632
- "step": 400
633
  },
634
  {
635
- "epoch": 7.9411764705882355,
636
- "grad_norm": 0.8984375,
637
- "learning_rate": 2.4731465338547556e-05,
638
- "loss": 0.3029,
639
- "step": 405
640
  },
641
  {
642
  "epoch": 8.0,
643
- "eval_loss": 4.195121765136719,
644
- "eval_runtime": 2.0472,
645
- "eval_samples_per_second": 4.885,
646
- "eval_steps_per_second": 0.977,
647
- "step": 408
648
- },
649
- {
650
- "epoch": 8.03921568627451,
651
- "grad_norm": 0.796875,
652
- "learning_rate": 2.252285755095652e-05,
653
- "loss": 0.2897,
654
- "step": 410
655
- },
656
- {
657
- "epoch": 8.137254901960784,
658
- "grad_norm": 0.80078125,
659
- "learning_rate": 2.0404978652089323e-05,
660
- "loss": 0.2649,
661
- "step": 415
662
- },
663
- {
664
- "epoch": 8.235294117647058,
665
- "grad_norm": 0.87890625,
666
- "learning_rate": 1.8380308764377842e-05,
667
- "loss": 0.2647,
668
- "step": 420
669
- },
670
- {
671
- "epoch": 8.333333333333334,
672
- "grad_norm": 0.828125,
673
- "learning_rate": 1.6451218858706374e-05,
674
- "loss": 0.2644,
675
- "step": 425
676
- },
677
- {
678
- "epoch": 8.431372549019608,
679
- "grad_norm": 0.77734375,
680
- "learning_rate": 1.4619967977908156e-05,
681
- "loss": 0.2709,
682
- "step": 430
683
- },
684
- {
685
- "epoch": 8.529411764705882,
686
- "grad_norm": 0.79296875,
687
- "learning_rate": 1.2888700591334223e-05,
688
- "loss": 0.2621,
689
- "step": 435
690
  },
691
  {
692
- "epoch": 8.627450980392156,
693
- "grad_norm": 0.8125,
694
- "learning_rate": 1.1259444083592585e-05,
695
- "loss": 0.2628,
696
- "step": 440
697
- },
698
- {
699
- "epoch": 8.72549019607843,
700
- "grad_norm": 0.81640625,
701
- "learning_rate": 9.734106380398023e-06,
702
- "loss": 0.2556,
703
- "step": 445
704
  },
705
  {
706
- "epoch": 8.823529411764707,
707
- "grad_norm": 0.765625,
708
- "learning_rate": 8.314473714313719e-06,
709
- "loss": 0.2611,
710
- "step": 450
711
  },
712
  {
713
- "epoch": 8.92156862745098,
714
- "grad_norm": 0.80859375,
715
- "learning_rate": 7.002208532999932e-06,
716
- "loss": 0.259,
717
- "step": 455
718
  },
719
  {
720
  "epoch": 9.0,
721
- "eval_loss": 4.499588966369629,
722
- "eval_runtime": 2.0494,
723
- "eval_samples_per_second": 4.88,
724
- "eval_steps_per_second": 0.976,
725
- "step": 459
726
- },
727
- {
728
- "epoch": 9.019607843137255,
729
- "grad_norm": 0.765625,
730
- "learning_rate": 5.798847552420183e-06,
731
- "loss": 0.2564,
732
- "step": 460
733
- },
734
- {
735
- "epoch": 9.117647058823529,
736
- "grad_norm": 0.74609375,
737
- "learning_rate": 4.705799957284351e-06,
738
- "loss": 0.2573,
739
- "step": 465
740
- },
741
- {
742
- "epoch": 9.215686274509803,
743
- "grad_norm": 0.75,
744
- "learning_rate": 3.724345750835878e-06,
745
- "loss": 0.2564,
746
- "step": 470
747
- },
748
- {
749
- "epoch": 9.313725490196079,
750
- "grad_norm": 0.78515625,
751
- "learning_rate": 2.855634255915951e-06,
752
- "loss": 0.2487,
753
- "step": 475
754
  },
755
  {
756
- "epoch": 9.411764705882353,
757
- "grad_norm": 0.7421875,
758
- "learning_rate": 2.100682769059548e-06,
759
- "loss": 0.2552,
760
- "step": 480
761
- },
762
- {
763
- "epoch": 9.509803921568627,
764
- "grad_norm": 0.765625,
765
- "learning_rate": 1.4603753691998735e-06,
766
- "loss": 0.253,
767
- "step": 485
768
- },
769
- {
770
- "epoch": 9.607843137254902,
771
- "grad_norm": 0.7578125,
772
- "learning_rate": 9.354618823758655e-07,
773
- "loss": 0.251,
774
- "step": 490
775
- },
776
- {
777
- "epoch": 9.705882352941176,
778
- "grad_norm": 0.734375,
779
- "learning_rate": 5.265570036553813e-07,
780
- "loss": 0.2528,
781
- "step": 495
782
- },
783
- {
784
- "epoch": 9.803921568627452,
785
- "grad_norm": 0.76171875,
786
- "learning_rate": 2.3413957730226144e-07,
787
- "loss": 0.2577,
788
- "step": 500
789
  },
790
  {
791
- "epoch": 9.901960784313726,
792
- "grad_norm": 0.7890625,
793
- "learning_rate": 5.8552036030179445e-08,
794
- "loss": 0.2508,
795
- "step": 505
796
  },
797
  {
798
  "epoch": 10.0,
799
- "grad_norm": 0.76953125,
800
  "learning_rate": 0.0,
801
- "loss": 0.2532,
802
- "step": 510
803
  },
804
  {
805
  "epoch": 10.0,
806
- "eval_loss": 4.513559818267822,
807
- "eval_runtime": 2.0473,
808
- "eval_samples_per_second": 4.884,
809
- "eval_steps_per_second": 0.977,
810
- "step": 510
811
  },
812
  {
813
  "epoch": 10.0,
814
- "step": 510,
815
- "total_flos": 3.8986265891805594e+17,
816
- "train_loss": 2.3579357918571024,
817
- "train_runtime": 3338.7641,
818
- "train_samples_per_second": 2.441,
819
- "train_steps_per_second": 0.153
820
  }
821
  ],
822
  "logging_steps": 5,
823
- "max_steps": 510,
824
  "num_input_tokens_seen": 0,
825
  "num_train_epochs": 10,
826
  "save_steps": 100,
827
- "total_flos": 3.8986265891805594e+17,
828
  "train_batch_size": 4,
829
  "trial_name": null,
830
  "trial_params": null
 
3
  "best_model_checkpoint": null,
4
  "epoch": 10.0,
5
  "eval_steps": 500,
6
+ "global_step": 140,
7
  "is_hyper_param_search": false,
8
  "is_local_process_zero": true,
9
  "is_world_process_zero": true,
10
  "log_history": [
11
  {
12
+ "epoch": 0.07142857142857142,
13
+ "grad_norm": 179.0,
14
+ "learning_rate": 1.4285714285714285e-05,
15
+ "loss": 46.5124,
16
  "step": 1
17
  },
18
  {
19
+ "epoch": 0.35714285714285715,
20
+ "grad_norm": 134.0,
21
+ "learning_rate": 7.142857142857143e-05,
22
+ "loss": 45.4999,
23
  "step": 5
24
  },
25
  {
26
+ "epoch": 0.7142857142857143,
27
+ "grad_norm": 19.5,
28
+ "learning_rate": 0.00014285714285714287,
29
+ "loss": 30.8653,
30
  "step": 10
31
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
32
  {
33
  "epoch": 1.0,
34
+ "eval_loss": 10.663816452026367,
35
+ "eval_runtime": 0.2491,
36
+ "eval_samples_per_second": 40.143,
37
+ "eval_steps_per_second": 4.014,
38
+ "step": 14
39
  },
40
  {
41
+ "epoch": 1.0714285714285714,
42
+ "grad_norm": 12.125,
43
+ "learning_rate": 0.00019996891820008164,
44
+ "loss": 22.9738,
45
+ "step": 15
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
46
  },
47
  {
48
+ "epoch": 1.4285714285714286,
49
+ "grad_norm": 4.53125,
50
+ "learning_rate": 0.00019888308262251285,
51
+ "loss": 19.8669,
52
+ "step": 20
53
  },
54
  {
55
+ "epoch": 1.7857142857142856,
56
+ "grad_norm": 5.5625,
57
+ "learning_rate": 0.0001962624246950012,
58
+ "loss": 18.5328,
59
+ "step": 25
60
  },
61
  {
62
  "epoch": 2.0,
63
+ "eval_loss": 7.303110599517822,
64
+ "eval_runtime": 0.2339,
65
+ "eval_samples_per_second": 42.753,
66
+ "eval_steps_per_second": 4.275,
67
+ "step": 28
68
  },
69
  {
70
+ "epoch": 2.142857142857143,
71
+ "grad_norm": 10.25,
72
+ "learning_rate": 0.00019214762118704076,
73
+ "loss": 17.3052,
74
+ "step": 30
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75
  },
76
  {
77
+ "epoch": 2.5,
78
+ "grad_norm": 17.5,
79
+ "learning_rate": 0.00018660254037844388,
80
+ "loss": 15.1835,
81
+ "step": 35
82
  },
83
  {
84
+ "epoch": 2.857142857142857,
85
+ "grad_norm": 24.875,
86
+ "learning_rate": 0.00017971325072229226,
87
+ "loss": 11.486,
88
+ "step": 40
89
  },
90
  {
91
  "epoch": 3.0,
92
+ "eval_loss": 6.628046989440918,
93
+ "eval_runtime": 0.2415,
94
+ "eval_samples_per_second": 41.402,
95
+ "eval_steps_per_second": 4.14,
96
+ "step": 42
97
  },
98
  {
99
+ "epoch": 3.2142857142857144,
100
+ "grad_norm": 26.5,
101
+ "learning_rate": 0.00017158668492597186,
102
+ "loss": 6.607,
103
+ "step": 45
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
104
  },
105
  {
106
+ "epoch": 3.571428571428571,
107
+ "grad_norm": 7.125,
108
+ "learning_rate": 0.00016234898018587337,
109
+ "loss": 3.2178,
110
+ "step": 50
111
  },
112
  {
113
+ "epoch": 3.928571428571429,
114
+ "grad_norm": 5.4375,
115
+ "learning_rate": 0.0001521435203379498,
116
+ "loss": 2.4959,
117
+ "step": 55
118
  },
119
  {
120
  "epoch": 4.0,
121
+ "eval_loss": 3.508697032928467,
122
+ "eval_runtime": 0.2457,
123
+ "eval_samples_per_second": 40.698,
124
+ "eval_steps_per_second": 4.07,
125
+ "step": 56
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
126
  },
127
  {
128
+ "epoch": 4.285714285714286,
129
+ "grad_norm": 2.421875,
130
+ "learning_rate": 0.00014112871031306119,
131
+ "loss": 2.101,
132
+ "step": 60
133
  },
134
  {
135
+ "epoch": 4.642857142857143,
136
+ "grad_norm": 1.03125,
137
+ "learning_rate": 0.00012947551744109043,
138
+ "loss": 1.8643,
139
+ "step": 65
140
  },
141
  {
142
  "epoch": 5.0,
143
+ "grad_norm": 0.65234375,
144
  "learning_rate": 0.00011736481776669306,
145
+ "loss": 1.742,
146
+ "step": 70
147
  },
148
  {
149
  "epoch": 5.0,
150
+ "eval_loss": 3.0216174125671387,
151
+ "eval_runtime": 0.2345,
152
+ "eval_samples_per_second": 42.653,
153
+ "eval_steps_per_second": 4.265,
154
+ "step": 70
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
155
  },
156
  {
157
+ "epoch": 5.357142857142857,
158
+ "grad_norm": 0.75390625,
159
+ "learning_rate": 0.00010498458856606972,
160
+ "loss": 1.6462,
161
+ "step": 75
162
  },
163
  {
164
+ "epoch": 5.714285714285714,
165
+ "grad_norm": 0.73046875,
166
+ "learning_rate": 9.252699064135758e-05,
167
+ "loss": 1.5971,
168
+ "step": 80
169
  },
170
  {
171
  "epoch": 6.0,
172
+ "eval_loss": 2.880188465118408,
173
+ "eval_runtime": 0.2347,
174
+ "eval_samples_per_second": 42.612,
175
+ "eval_steps_per_second": 4.261,
176
+ "step": 84
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
177
  },
178
  {
179
+ "epoch": 6.071428571428571,
180
+ "grad_norm": 0.474609375,
181
+ "learning_rate": 8.018538568006027e-05,
182
+ "loss": 1.5391,
183
+ "step": 85
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
184
  },
185
  {
186
+ "epoch": 6.428571428571429,
187
+ "grad_norm": 0.62109375,
188
+ "learning_rate": 6.815133497483157e-05,
189
+ "loss": 1.5124,
190
+ "step": 90
191
  },
192
  {
193
+ "epoch": 6.785714285714286,
194
+ "grad_norm": 0.408203125,
195
+ "learning_rate": 5.6611626088244194e-05,
196
+ "loss": 1.4792,
197
+ "step": 95
198
  },
199
  {
200
  "epoch": 7.0,
201
+ "eval_loss": 2.8307275772094727,
202
+ "eval_runtime": 0.235,
203
+ "eval_samples_per_second": 42.562,
204
+ "eval_steps_per_second": 4.256,
205
+ "step": 98
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
206
  },
207
  {
208
+ "epoch": 7.142857142857143,
209
+ "grad_norm": 0.443359375,
210
+ "learning_rate": 4.574537361342407e-05,
211
+ "loss": 1.4514,
212
+ "step": 100
 
 
 
 
 
 
 
 
 
 
 
 
 
 
213
  },
214
  {
215
+ "epoch": 7.5,
216
+ "grad_norm": 0.60546875,
217
+ "learning_rate": 3.5721239031346066e-05,
218
+ "loss": 1.4383,
219
+ "step": 105
220
  },
221
  {
222
+ "epoch": 7.857142857142857,
223
+ "grad_norm": 0.8828125,
224
+ "learning_rate": 2.669481281701739e-05,
225
+ "loss": 1.4333,
226
+ "step": 110
227
  },
228
  {
229
  "epoch": 8.0,
230
+ "eval_loss": 2.808138608932495,
231
+ "eval_runtime": 0.2344,
232
+ "eval_samples_per_second": 42.659,
233
+ "eval_steps_per_second": 4.266,
234
+ "step": 112
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
235
  },
236
  {
237
+ "epoch": 8.214285714285714,
238
+ "grad_norm": 0.482421875,
239
+ "learning_rate": 1.880619942841435e-05,
240
+ "loss": 1.4232,
241
+ "step": 115
 
 
 
 
 
 
 
242
  },
243
  {
244
+ "epoch": 8.571428571428571,
245
+ "grad_norm": 0.78125,
246
+ "learning_rate": 1.2177842662977135e-05,
247
+ "loss": 1.4159,
248
+ "step": 120
249
  },
250
  {
251
+ "epoch": 8.928571428571429,
252
+ "grad_norm": 0.53515625,
253
+ "learning_rate": 6.9126251355795864e-06,
254
+ "loss": 1.4129,
255
+ "step": 125
256
  },
257
  {
258
  "epoch": 9.0,
259
+ "eval_loss": 2.8151421546936035,
260
+ "eval_runtime": 0.2356,
261
+ "eval_samples_per_second": 42.443,
262
+ "eval_steps_per_second": 4.244,
263
+ "step": 126
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
264
  },
265
  {
266
+ "epoch": 9.285714285714286,
267
+ "grad_norm": 0.427734375,
268
+ "learning_rate": 3.092271377092215e-06,
269
+ "loss": 1.4193,
270
+ "step": 130
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
271
  },
272
  {
273
+ "epoch": 9.642857142857142,
274
+ "grad_norm": 0.6875,
275
+ "learning_rate": 7.760793399827937e-07,
276
+ "loss": 1.4139,
277
+ "step": 135
278
  },
279
  {
280
  "epoch": 10.0,
281
+ "grad_norm": 0.376953125,
282
  "learning_rate": 0.0,
283
+ "loss": 1.4048,
284
+ "step": 140
285
  },
286
  {
287
  "epoch": 10.0,
288
+ "eval_loss": 2.812880277633667,
289
+ "eval_runtime": 0.2405,
290
+ "eval_samples_per_second": 41.578,
291
+ "eval_steps_per_second": 4.158,
292
+ "step": 140
293
  },
294
  {
295
  "epoch": 10.0,
296
+ "step": 140,
297
+ "total_flos": 4.268849030789857e+17,
298
+ "train_loss": 7.876109651156834,
299
+ "train_runtime": 340.0833,
300
+ "train_samples_per_second": 25.758,
301
+ "train_steps_per_second": 0.412
302
  }
303
  ],
304
  "logging_steps": 5,
305
+ "max_steps": 140,
306
  "num_input_tokens_seen": 0,
307
  "num_train_epochs": 10,
308
  "save_steps": 100,
309
+ "total_flos": 4.268849030789857e+17,
310
  "train_batch_size": 4,
311
  "trial_name": null,
312
  "trial_params": null
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:090af3d5439a52bc14b85b7cae16a338c260d9008f2e037d831fd214739b3782
3
  size 5176
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0007bf634ec1be91eaf557c03316891e1efffff59f85ff61b4e10440de3d63bd
3
  size 5176