Model save
Browse files- README.md +57 -0
- all_results.json +9 -0
- train_results.json +9 -0
- trainer_state.json +253 -0
README.md
ADDED
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: mistralai/Mistral-7B-Instruct-v0.2
|
3 |
+
library_name: transformers
|
4 |
+
model_name: prometheus-7b-v1.5-beta-1-0109-sum-pref-overfit-v3-flash
|
5 |
+
tags:
|
6 |
+
- generated_from_trainer
|
7 |
+
- trl
|
8 |
+
- sft
|
9 |
+
licence: license
|
10 |
+
---
|
11 |
+
|
12 |
+
# Model Card for prometheus-7b-v1.5-beta-1-0109-sum-pref-overfit-v3-flash
|
13 |
+
|
14 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-Instruct-v0.2](https://huggingface.co/mistralai/Mistral-7B-Instruct-v0.2).
|
15 |
+
It has been trained using [TRL](https://github.com/huggingface/trl).
|
16 |
+
|
17 |
+
## Quick start
|
18 |
+
|
19 |
+
```python
|
20 |
+
from transformers import pipeline
|
21 |
+
|
22 |
+
question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
|
23 |
+
generator = pipeline("text-generation", model="christineyu/prometheus-7b-v1.5-beta-1-0109-sum-pref-overfit-v3-flash", device="cuda")
|
24 |
+
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
|
25 |
+
print(output["generated_text"])
|
26 |
+
```
|
27 |
+
|
28 |
+
## Training procedure
|
29 |
+
|
30 |
+
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://wandb.ai/myexp/huggingface/runs/7ncrpm53)
|
31 |
+
|
32 |
+
This model was trained with SFT.
|
33 |
+
|
34 |
+
### Framework versions
|
35 |
+
|
36 |
+
- TRL: 0.12.2
|
37 |
+
- Transformers: 4.46.3
|
38 |
+
- Pytorch: 2.3.0
|
39 |
+
- Datasets: 3.2.0
|
40 |
+
- Tokenizers: 0.20.3
|
41 |
+
|
42 |
+
## Citations
|
43 |
+
|
44 |
+
|
45 |
+
|
46 |
+
Cite TRL as:
|
47 |
+
|
48 |
+
```bibtex
|
49 |
+
@misc{vonwerra2022trl,
|
50 |
+
title = {{TRL: Transformer Reinforcement Learning}},
|
51 |
+
author = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
|
52 |
+
year = 2020,
|
53 |
+
journal = {GitHub repository},
|
54 |
+
publisher = {GitHub},
|
55 |
+
howpublished = {\url{https://github.com/huggingface/trl}}
|
56 |
+
}
|
57 |
+
```
|
all_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"total_flos": 2.5427771438137344e+16,
|
4 |
+
"train_loss": 0.06073189970825845,
|
5 |
+
"train_runtime": 598.238,
|
6 |
+
"train_samples": 990,
|
7 |
+
"train_samples_per_second": 0.236,
|
8 |
+
"train_steps_per_second": 0.236
|
9 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"total_flos": 2.5427771438137344e+16,
|
4 |
+
"train_loss": 0.06073189970825845,
|
5 |
+
"train_runtime": 598.238,
|
6 |
+
"train_samples": 990,
|
7 |
+
"train_samples_per_second": 0.236,
|
8 |
+
"train_steps_per_second": 0.236
|
9 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,253 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 141,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.0070921985815602835,
|
13 |
+
"grad_norm": 8.373100280761719,
|
14 |
+
"learning_rate": 6.666666666666667e-07,
|
15 |
+
"loss": 0.5467,
|
16 |
+
"step": 1
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.03546099290780142,
|
20 |
+
"grad_norm": 5.184474945068359,
|
21 |
+
"learning_rate": 3.3333333333333333e-06,
|
22 |
+
"loss": 0.5472,
|
23 |
+
"step": 5
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.07092198581560284,
|
27 |
+
"grad_norm": 2.155921697616577,
|
28 |
+
"learning_rate": 6.666666666666667e-06,
|
29 |
+
"loss": 0.4131,
|
30 |
+
"step": 10
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.10638297872340426,
|
34 |
+
"grad_norm": 1.2390508651733398,
|
35 |
+
"learning_rate": 1e-05,
|
36 |
+
"loss": 0.291,
|
37 |
+
"step": 15
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.14184397163120568,
|
41 |
+
"grad_norm": 1.1246978044509888,
|
42 |
+
"learning_rate": 9.961196033000862e-06,
|
43 |
+
"loss": 0.1965,
|
44 |
+
"step": 20
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.1773049645390071,
|
48 |
+
"grad_norm": 1.2201730012893677,
|
49 |
+
"learning_rate": 9.84538643114539e-06,
|
50 |
+
"loss": 0.101,
|
51 |
+
"step": 25
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.2127659574468085,
|
55 |
+
"grad_norm": 2.5533862113952637,
|
56 |
+
"learning_rate": 9.654368743221022e-06,
|
57 |
+
"loss": 0.0424,
|
58 |
+
"step": 30
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.24822695035460993,
|
62 |
+
"grad_norm": 0.4596536457538605,
|
63 |
+
"learning_rate": 9.391107866851143e-06,
|
64 |
+
"loss": 0.0168,
|
65 |
+
"step": 35
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.28368794326241137,
|
69 |
+
"grad_norm": 0.3188921809196472,
|
70 |
+
"learning_rate": 9.059690028579285e-06,
|
71 |
+
"loss": 0.0103,
|
72 |
+
"step": 40
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.3191489361702128,
|
76 |
+
"grad_norm": 0.2514072060585022,
|
77 |
+
"learning_rate": 8.665259359149132e-06,
|
78 |
+
"loss": 0.0089,
|
79 |
+
"step": 45
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.3546099290780142,
|
83 |
+
"grad_norm": 4.992562294006348,
|
84 |
+
"learning_rate": 8.213938048432697e-06,
|
85 |
+
"loss": 0.0071,
|
86 |
+
"step": 50
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.3900709219858156,
|
90 |
+
"grad_norm": 0.2883487343788147,
|
91 |
+
"learning_rate": 7.712731319328798e-06,
|
92 |
+
"loss": 0.0063,
|
93 |
+
"step": 55
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.425531914893617,
|
97 |
+
"grad_norm": 0.8716890215873718,
|
98 |
+
"learning_rate": 7.169418695587791e-06,
|
99 |
+
"loss": 0.0076,
|
100 |
+
"step": 60
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.46099290780141844,
|
104 |
+
"grad_norm": 0.2935401201248169,
|
105 |
+
"learning_rate": 6.592433251258423e-06,
|
106 |
+
"loss": 0.007,
|
107 |
+
"step": 65
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.49645390070921985,
|
111 |
+
"grad_norm": 0.11323031038045883,
|
112 |
+
"learning_rate": 5.990730715996989e-06,
|
113 |
+
"loss": 0.0053,
|
114 |
+
"step": 70
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.5319148936170213,
|
118 |
+
"grad_norm": 0.45351043343544006,
|
119 |
+
"learning_rate": 5.373650467932122e-06,
|
120 |
+
"loss": 0.0044,
|
121 |
+
"step": 75
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.5673758865248227,
|
125 |
+
"grad_norm": 0.22218099236488342,
|
126 |
+
"learning_rate": 4.750770571696514e-06,
|
127 |
+
"loss": 0.0038,
|
128 |
+
"step": 80
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.6028368794326241,
|
132 |
+
"grad_norm": 0.1962103545665741,
|
133 |
+
"learning_rate": 4.131759111665349e-06,
|
134 |
+
"loss": 0.0032,
|
135 |
+
"step": 85
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.6382978723404256,
|
139 |
+
"grad_norm": 0.7398797273635864,
|
140 |
+
"learning_rate": 3.526224127945479e-06,
|
141 |
+
"loss": 0.0046,
|
142 |
+
"step": 90
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.6737588652482269,
|
146 |
+
"grad_norm": 0.21601063013076782,
|
147 |
+
"learning_rate": 2.9435644843469434e-06,
|
148 |
+
"loss": 0.0038,
|
149 |
+
"step": 95
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.7092198581560284,
|
153 |
+
"grad_norm": 0.13288450241088867,
|
154 |
+
"learning_rate": 2.39282398310251e-06,
|
155 |
+
"loss": 0.0036,
|
156 |
+
"step": 100
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.7446808510638298,
|
160 |
+
"grad_norm": 0.1287090927362442,
|
161 |
+
"learning_rate": 1.8825509907063328e-06,
|
162 |
+
"loss": 0.0037,
|
163 |
+
"step": 105
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.7801418439716312,
|
167 |
+
"grad_norm": 0.291775107383728,
|
168 |
+
"learning_rate": 1.4206657537014078e-06,
|
169 |
+
"loss": 0.003,
|
170 |
+
"step": 110
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.8156028368794326,
|
174 |
+
"grad_norm": 0.2999476194381714,
|
175 |
+
"learning_rate": 1.0143374638853892e-06,
|
176 |
+
"loss": 0.0042,
|
177 |
+
"step": 115
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.851063829787234,
|
181 |
+
"grad_norm": 0.2297772467136383,
|
182 |
+
"learning_rate": 6.698729810778065e-07,
|
183 |
+
"loss": 0.0041,
|
184 |
+
"step": 120
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.8865248226950354,
|
188 |
+
"grad_norm": 0.1276993751525879,
|
189 |
+
"learning_rate": 3.9261894064796136e-07,
|
190 |
+
"loss": 0.0029,
|
191 |
+
"step": 125
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.9219858156028369,
|
195 |
+
"grad_norm": 0.2057814598083496,
|
196 |
+
"learning_rate": 1.8687876524993987e-07,
|
197 |
+
"loss": 0.0049,
|
198 |
+
"step": 130
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.9574468085106383,
|
202 |
+
"grad_norm": 0.33941203355789185,
|
203 |
+
"learning_rate": 5.584586887435739e-08,
|
204 |
+
"loss": 0.0032,
|
205 |
+
"step": 135
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.9929078014184397,
|
209 |
+
"grad_norm": 0.39472323656082153,
|
210 |
+
"learning_rate": 1.5540899959187727e-09,
|
211 |
+
"loss": 0.0026,
|
212 |
+
"step": 140
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 1.0,
|
216 |
+
"eval_loss": 0.010593634098768234,
|
217 |
+
"eval_runtime": 1.2814,
|
218 |
+
"eval_samples_per_second": 0.78,
|
219 |
+
"eval_steps_per_second": 0.78,
|
220 |
+
"step": 141
|
221 |
+
},
|
222 |
+
{
|
223 |
+
"epoch": 1.0,
|
224 |
+
"step": 141,
|
225 |
+
"total_flos": 2.5427771438137344e+16,
|
226 |
+
"train_loss": 0.06073189970825845,
|
227 |
+
"train_runtime": 598.238,
|
228 |
+
"train_samples_per_second": 0.236,
|
229 |
+
"train_steps_per_second": 0.236
|
230 |
+
}
|
231 |
+
],
|
232 |
+
"logging_steps": 5,
|
233 |
+
"max_steps": 141,
|
234 |
+
"num_input_tokens_seen": 0,
|
235 |
+
"num_train_epochs": 1,
|
236 |
+
"save_steps": 100,
|
237 |
+
"stateful_callbacks": {
|
238 |
+
"TrainerControl": {
|
239 |
+
"args": {
|
240 |
+
"should_epoch_stop": false,
|
241 |
+
"should_evaluate": false,
|
242 |
+
"should_log": false,
|
243 |
+
"should_save": true,
|
244 |
+
"should_training_stop": true
|
245 |
+
},
|
246 |
+
"attributes": {}
|
247 |
+
}
|
248 |
+
},
|
249 |
+
"total_flos": 2.5427771438137344e+16,
|
250 |
+
"train_batch_size": 1,
|
251 |
+
"trial_name": null,
|
252 |
+
"trial_params": null
|
253 |
+
}
|