AmberYifan
commited on
Commit
•
f39d3a4
1
Parent(s):
f527e33
Model save
Browse files- README.md +66 -0
- all_results.json +9 -0
- generation_config.json +6 -0
- train_results.json +9 -0
- trainer_state.json +316 -0
README.md
ADDED
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
base_model: mistralai/Mistral-7B-v0.1
|
4 |
+
tags:
|
5 |
+
- trl
|
6 |
+
- sft
|
7 |
+
- generated_from_trainer
|
8 |
+
datasets:
|
9 |
+
- generator
|
10 |
+
model-index:
|
11 |
+
- name: mistral-safe-sft-full
|
12 |
+
results: []
|
13 |
+
---
|
14 |
+
|
15 |
+
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
16 |
+
should probably proofread and complete it, then remove this comment. -->
|
17 |
+
|
18 |
+
# mistral-safe-sft-full
|
19 |
+
|
20 |
+
This model is a fine-tuned version of [mistralai/Mistral-7B-v0.1](https://huggingface.co/mistralai/Mistral-7B-v0.1) on the generator dataset.
|
21 |
+
It achieves the following results on the evaluation set:
|
22 |
+
- Loss: 1.2626
|
23 |
+
|
24 |
+
## Model description
|
25 |
+
|
26 |
+
More information needed
|
27 |
+
|
28 |
+
## Intended uses & limitations
|
29 |
+
|
30 |
+
More information needed
|
31 |
+
|
32 |
+
## Training and evaluation data
|
33 |
+
|
34 |
+
More information needed
|
35 |
+
|
36 |
+
## Training procedure
|
37 |
+
|
38 |
+
### Training hyperparameters
|
39 |
+
|
40 |
+
The following hyperparameters were used during training:
|
41 |
+
- learning_rate: 2e-05
|
42 |
+
- train_batch_size: 8
|
43 |
+
- eval_batch_size: 8
|
44 |
+
- seed: 42
|
45 |
+
- distributed_type: multi-GPU
|
46 |
+
- num_devices: 4
|
47 |
+
- total_train_batch_size: 32
|
48 |
+
- total_eval_batch_size: 32
|
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
|
52 |
+
- num_epochs: 1
|
53 |
+
|
54 |
+
### Training results
|
55 |
+
|
56 |
+
| Training Loss | Epoch | Step | Validation Loss |
|
57 |
+
|:-------------:|:-----:|:----:|:---------------:|
|
58 |
+
| 1.2583 | 1.0 | 185 | 1.2626 |
|
59 |
+
|
60 |
+
|
61 |
+
### Framework versions
|
62 |
+
|
63 |
+
- Transformers 4.43.3
|
64 |
+
- Pytorch 2.2.2+cu121
|
65 |
+
- Datasets 2.20.0
|
66 |
+
- Tokenizers 0.19.1
|
all_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"total_flos": 19367618150400.0,
|
4 |
+
"train_loss": 1.3619260220914273,
|
5 |
+
"train_runtime": 1878.7186,
|
6 |
+
"train_samples": 75077,
|
7 |
+
"train_samples_per_second": 3.148,
|
8 |
+
"train_steps_per_second": 0.098
|
9 |
+
}
|
generation_config.json
ADDED
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"_from_model_config": true,
|
3 |
+
"bos_token_id": 1,
|
4 |
+
"eos_token_id": 2,
|
5 |
+
"transformers_version": "4.43.3"
|
6 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 1.0,
|
3 |
+
"total_flos": 19367618150400.0,
|
4 |
+
"train_loss": 1.3619260220914273,
|
5 |
+
"train_runtime": 1878.7186,
|
6 |
+
"train_samples": 75077,
|
7 |
+
"train_samples_per_second": 3.148,
|
8 |
+
"train_steps_per_second": 0.098
|
9 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,316 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 1.0,
|
5 |
+
"eval_steps": 500,
|
6 |
+
"global_step": 185,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.005405405405405406,
|
13 |
+
"grad_norm": 61.001103964725324,
|
14 |
+
"learning_rate": 1.0526315789473685e-06,
|
15 |
+
"loss": 1.9097,
|
16 |
+
"step": 1
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.02702702702702703,
|
20 |
+
"grad_norm": 35.97997402423919,
|
21 |
+
"learning_rate": 5.263157894736842e-06,
|
22 |
+
"loss": 1.7102,
|
23 |
+
"step": 5
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.05405405405405406,
|
27 |
+
"grad_norm": 30.10156193198182,
|
28 |
+
"learning_rate": 1.0526315789473684e-05,
|
29 |
+
"loss": 1.5516,
|
30 |
+
"step": 10
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.08108108108108109,
|
34 |
+
"grad_norm": 6.532111146208257,
|
35 |
+
"learning_rate": 1.578947368421053e-05,
|
36 |
+
"loss": 1.4536,
|
37 |
+
"step": 15
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.10810810810810811,
|
41 |
+
"grad_norm": 4.983904851506608,
|
42 |
+
"learning_rate": 1.999820922669738e-05,
|
43 |
+
"loss": 1.4112,
|
44 |
+
"step": 20
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.13513513513513514,
|
48 |
+
"grad_norm": 3.8732866550483545,
|
49 |
+
"learning_rate": 1.993559947963185e-05,
|
50 |
+
"loss": 1.4056,
|
51 |
+
"step": 25
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.16216216216216217,
|
55 |
+
"grad_norm": 3.0620367608079846,
|
56 |
+
"learning_rate": 1.9784091409455728e-05,
|
57 |
+
"loss": 1.4016,
|
58 |
+
"step": 30
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.1891891891891892,
|
62 |
+
"grad_norm": 3.2352358324557167,
|
63 |
+
"learning_rate": 1.9545040627715554e-05,
|
64 |
+
"loss": 1.4128,
|
65 |
+
"step": 35
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.21621621621621623,
|
69 |
+
"grad_norm": 3.022703286965415,
|
70 |
+
"learning_rate": 1.9220586030376135e-05,
|
71 |
+
"loss": 1.4148,
|
72 |
+
"step": 40
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 0.24324324324324326,
|
76 |
+
"grad_norm": 2.8955611484982033,
|
77 |
+
"learning_rate": 1.881363066014649e-05,
|
78 |
+
"loss": 1.4159,
|
79 |
+
"step": 45
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 0.2702702702702703,
|
83 |
+
"grad_norm": 2.8660620524615688,
|
84 |
+
"learning_rate": 1.8327815731637612e-05,
|
85 |
+
"loss": 1.42,
|
86 |
+
"step": 50
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 0.2972972972972973,
|
90 |
+
"grad_norm": 2.632475843797219,
|
91 |
+
"learning_rate": 1.7767488051760858e-05,
|
92 |
+
"loss": 1.4078,
|
93 |
+
"step": 55
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 0.32432432432432434,
|
97 |
+
"grad_norm": 2.5436272914892872,
|
98 |
+
"learning_rate": 1.713766112687139e-05,
|
99 |
+
"loss": 1.4015,
|
100 |
+
"step": 60
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 0.35135135135135137,
|
104 |
+
"grad_norm": 2.431239149211266,
|
105 |
+
"learning_rate": 1.644397030464877e-05,
|
106 |
+
"loss": 1.3807,
|
107 |
+
"step": 65
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 0.3783783783783784,
|
111 |
+
"grad_norm": 2.328539582829742,
|
112 |
+
"learning_rate": 1.5692622352080662e-05,
|
113 |
+
"loss": 1.3639,
|
114 |
+
"step": 70
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 0.40540540540540543,
|
118 |
+
"grad_norm": 2.708625235728292,
|
119 |
+
"learning_rate": 1.4890339920698334e-05,
|
120 |
+
"loss": 1.3836,
|
121 |
+
"step": 75
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 0.43243243243243246,
|
125 |
+
"grad_norm": 2.4202434897618317,
|
126 |
+
"learning_rate": 1.404430139595877e-05,
|
127 |
+
"loss": 1.3823,
|
128 |
+
"step": 80
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 0.4594594594594595,
|
132 |
+
"grad_norm": 2.243346482513894,
|
133 |
+
"learning_rate": 1.316207666896824e-05,
|
134 |
+
"loss": 1.3538,
|
135 |
+
"step": 85
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 0.4864864864864865,
|
139 |
+
"grad_norm": 2.4109049967489433,
|
140 |
+
"learning_rate": 1.2251559405226943e-05,
|
141 |
+
"loss": 1.3503,
|
142 |
+
"step": 90
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 0.5135135135135135,
|
146 |
+
"grad_norm": 2.242883116591874,
|
147 |
+
"learning_rate": 1.1320896416417026e-05,
|
148 |
+
"loss": 1.3562,
|
149 |
+
"step": 95
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 0.5405405405405406,
|
153 |
+
"grad_norm": 2.1959662600808203,
|
154 |
+
"learning_rate": 1.0378414767176706e-05,
|
155 |
+
"loss": 1.3447,
|
156 |
+
"step": 100
|
157 |
+
},
|
158 |
+
{
|
159 |
+
"epoch": 0.5675675675675675,
|
160 |
+
"grad_norm": 2.2348646553535714,
|
161 |
+
"learning_rate": 9.43254726906926e-06,
|
162 |
+
"loss": 1.3398,
|
163 |
+
"step": 105
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 0.5945945945945946,
|
167 |
+
"grad_norm": 2.21405256524929,
|
168 |
+
"learning_rate": 8.491757028386262e-06,
|
169 |
+
"loss": 1.3383,
|
170 |
+
"step": 110
|
171 |
+
},
|
172 |
+
{
|
173 |
+
"epoch": 0.6216216216216216,
|
174 |
+
"grad_norm": 2.1710850419967263,
|
175 |
+
"learning_rate": 7.564461722890082e-06,
|
176 |
+
"loss": 1.3271,
|
177 |
+
"step": 115
|
178 |
+
},
|
179 |
+
{
|
180 |
+
"epoch": 0.6486486486486487,
|
181 |
+
"grad_norm": 2.1752417729097338,
|
182 |
+
"learning_rate": 6.6589582850261025e-06,
|
183 |
+
"loss": 1.3176,
|
184 |
+
"step": 120
|
185 |
+
},
|
186 |
+
{
|
187 |
+
"epoch": 0.6756756756756757,
|
188 |
+
"grad_norm": 2.1322248363189686,
|
189 |
+
"learning_rate": 5.78334866549816e-06,
|
190 |
+
"loss": 1.321,
|
191 |
+
"step": 125
|
192 |
+
},
|
193 |
+
{
|
194 |
+
"epoch": 0.7027027027027027,
|
195 |
+
"grad_norm": 2.1516864548066525,
|
196 |
+
"learning_rate": 4.9454673414341945e-06,
|
197 |
+
"loss": 1.3074,
|
198 |
+
"step": 130
|
199 |
+
},
|
200 |
+
{
|
201 |
+
"epoch": 0.7297297297297297,
|
202 |
+
"grad_norm": 2.04797883118049,
|
203 |
+
"learning_rate": 4.152811217759529e-06,
|
204 |
+
"loss": 1.3123,
|
205 |
+
"step": 135
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 0.7567567567567568,
|
209 |
+
"grad_norm": 2.1202023783270367,
|
210 |
+
"learning_rate": 3.4124725489820643e-06,
|
211 |
+
"loss": 1.3026,
|
212 |
+
"step": 140
|
213 |
+
},
|
214 |
+
{
|
215 |
+
"epoch": 0.7837837837837838,
|
216 |
+
"grad_norm": 2.1203746539007158,
|
217 |
+
"learning_rate": 2.7310754815685627e-06,
|
218 |
+
"loss": 1.2838,
|
219 |
+
"step": 145
|
220 |
+
},
|
221 |
+
{
|
222 |
+
"epoch": 0.8108108108108109,
|
223 |
+
"grad_norm": 2.058016060238832,
|
224 |
+
"learning_rate": 2.114716784696342e-06,
|
225 |
+
"loss": 1.2805,
|
226 |
+
"step": 150
|
227 |
+
},
|
228 |
+
{
|
229 |
+
"epoch": 0.8378378378378378,
|
230 |
+
"grad_norm": 2.072702697896577,
|
231 |
+
"learning_rate": 1.5689112996891576e-06,
|
232 |
+
"loss": 1.276,
|
233 |
+
"step": 155
|
234 |
+
},
|
235 |
+
{
|
236 |
+
"epoch": 0.8648648648648649,
|
237 |
+
"grad_norm": 2.070358297476048,
|
238 |
+
"learning_rate": 1.0985425962260342e-06,
|
239 |
+
"loss": 1.2768,
|
240 |
+
"step": 160
|
241 |
+
},
|
242 |
+
{
|
243 |
+
"epoch": 0.8918918918918919,
|
244 |
+
"grad_norm": 1.993421902652665,
|
245 |
+
"learning_rate": 7.078192768243486e-07,
|
246 |
+
"loss": 1.2755,
|
247 |
+
"step": 165
|
248 |
+
},
|
249 |
+
{
|
250 |
+
"epoch": 0.918918918918919,
|
251 |
+
"grad_norm": 2.02321205406663,
|
252 |
+
"learning_rate": 4.0023732056077235e-07,
|
253 |
+
"loss": 1.276,
|
254 |
+
"step": 170
|
255 |
+
},
|
256 |
+
{
|
257 |
+
"epoch": 0.9459459459459459,
|
258 |
+
"grad_norm": 2.0216958269361154,
|
259 |
+
"learning_rate": 1.7854880295797406e-07,
|
260 |
+
"loss": 1.2694,
|
261 |
+
"step": 175
|
262 |
+
},
|
263 |
+
{
|
264 |
+
"epoch": 0.972972972972973,
|
265 |
+
"grad_norm": 2.0002791761840157,
|
266 |
+
"learning_rate": 4.473727191441124e-08,
|
267 |
+
"loss": 1.2669,
|
268 |
+
"step": 180
|
269 |
+
},
|
270 |
+
{
|
271 |
+
"epoch": 1.0,
|
272 |
+
"grad_norm": 1.9847900492979156,
|
273 |
+
"learning_rate": 0.0,
|
274 |
+
"loss": 1.2583,
|
275 |
+
"step": 185
|
276 |
+
},
|
277 |
+
{
|
278 |
+
"epoch": 1.0,
|
279 |
+
"eval_loss": 1.2626361846923828,
|
280 |
+
"eval_runtime": 57.4547,
|
281 |
+
"eval_samples_per_second": 11.47,
|
282 |
+
"eval_steps_per_second": 0.366,
|
283 |
+
"step": 185
|
284 |
+
},
|
285 |
+
{
|
286 |
+
"epoch": 1.0,
|
287 |
+
"step": 185,
|
288 |
+
"total_flos": 19367618150400.0,
|
289 |
+
"train_loss": 1.3619260220914273,
|
290 |
+
"train_runtime": 1878.7186,
|
291 |
+
"train_samples_per_second": 3.148,
|
292 |
+
"train_steps_per_second": 0.098
|
293 |
+
}
|
294 |
+
],
|
295 |
+
"logging_steps": 5,
|
296 |
+
"max_steps": 185,
|
297 |
+
"num_input_tokens_seen": 0,
|
298 |
+
"num_train_epochs": 1,
|
299 |
+
"save_steps": 800,
|
300 |
+
"stateful_callbacks": {
|
301 |
+
"TrainerControl": {
|
302 |
+
"args": {
|
303 |
+
"should_epoch_stop": false,
|
304 |
+
"should_evaluate": false,
|
305 |
+
"should_log": false,
|
306 |
+
"should_save": true,
|
307 |
+
"should_training_stop": true
|
308 |
+
},
|
309 |
+
"attributes": {}
|
310 |
+
}
|
311 |
+
},
|
312 |
+
"total_flos": 19367618150400.0,
|
313 |
+
"train_batch_size": 8,
|
314 |
+
"trial_name": null,
|
315 |
+
"trial_params": null
|
316 |
+
}
|