hvgg1ngface
commited on
Commit
•
6b53d36
1
Parent(s):
d773fa2
End of training
Browse files- README.md +3 -2
- all_results.json +12 -0
- eval_results.json +7 -0
- train_results.json +8 -0
- trainer_state.json +309 -0
- training_eval_loss.png +0 -0
- training_loss.png +0 -0
README.md
CHANGED
@@ -4,6 +4,7 @@ library_name: peft
|
|
4 |
license: llama3
|
5 |
tags:
|
6 |
- llama-factory
|
|
|
7 |
- generated_from_trainer
|
8 |
model-index:
|
9 |
- name: wenjie_6k
|
@@ -15,9 +16,9 @@ should probably proofread and complete it, then remove this comment. -->
|
|
15 |
|
16 |
# wenjie_6k
|
17 |
|
18 |
-
This model is a fine-tuned version of [shenzhi-wang/Llama3-8B-Chinese-Chat](https://huggingface.co/shenzhi-wang/Llama3-8B-Chinese-Chat) on the
|
19 |
It achieves the following results on the evaluation set:
|
20 |
-
- Loss: 0.
|
21 |
|
22 |
## Model description
|
23 |
|
|
|
4 |
license: llama3
|
5 |
tags:
|
6 |
- llama-factory
|
7 |
+
- lora
|
8 |
- generated_from_trainer
|
9 |
model-index:
|
10 |
- name: wenjie_6k
|
|
|
16 |
|
17 |
# wenjie_6k
|
18 |
|
19 |
+
This model is a fine-tuned version of [shenzhi-wang/Llama3-8B-Chinese-Chat](https://huggingface.co/shenzhi-wang/Llama3-8B-Chinese-Chat) on the data_large2, the target_QA_short and the append_QA datasets.
|
20 |
It achieves the following results on the evaluation set:
|
21 |
+
- Loss: 0.9191
|
22 |
|
23 |
## Model description
|
24 |
|
all_results.json
ADDED
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.9892761394101877,
|
3 |
+
"eval_loss": 0.91908198595047,
|
4 |
+
"eval_runtime": 5.4522,
|
5 |
+
"eval_samples_per_second": 57.592,
|
6 |
+
"eval_steps_per_second": 1.834,
|
7 |
+
"total_flos": 3.7910142911093146e+17,
|
8 |
+
"train_loss": 1.1117366898444392,
|
9 |
+
"train_runtime": 1080.8388,
|
10 |
+
"train_samples_per_second": 22.064,
|
11 |
+
"train_steps_per_second": 0.344
|
12 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,7 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.9892761394101877,
|
3 |
+
"eval_loss": 0.91908198595047,
|
4 |
+
"eval_runtime": 5.4522,
|
5 |
+
"eval_samples_per_second": 57.592,
|
6 |
+
"eval_steps_per_second": 1.834
|
7 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 3.9892761394101877,
|
3 |
+
"total_flos": 3.7910142911093146e+17,
|
4 |
+
"train_loss": 1.1117366898444392,
|
5 |
+
"train_runtime": 1080.8388,
|
6 |
+
"train_samples_per_second": 22.064,
|
7 |
+
"train_steps_per_second": 0.344
|
8 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,309 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 3.9892761394101877,
|
5 |
+
"eval_steps": 200,
|
6 |
+
"global_step": 372,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.10723860589812333,
|
13 |
+
"grad_norm": 0.8259232640266418,
|
14 |
+
"learning_rate": 4.210526315789473e-06,
|
15 |
+
"loss": 1.7013,
|
16 |
+
"step": 10
|
17 |
+
},
|
18 |
+
{
|
19 |
+
"epoch": 0.21447721179624665,
|
20 |
+
"grad_norm": 0.8551827073097229,
|
21 |
+
"learning_rate": 7.999841591871243e-06,
|
22 |
+
"loss": 1.676,
|
23 |
+
"step": 20
|
24 |
+
},
|
25 |
+
{
|
26 |
+
"epoch": 0.32171581769436997,
|
27 |
+
"grad_norm": 1.056661605834961,
|
28 |
+
"learning_rate": 7.9808477930786e-06,
|
29 |
+
"loss": 1.6483,
|
30 |
+
"step": 30
|
31 |
+
},
|
32 |
+
{
|
33 |
+
"epoch": 0.4289544235924933,
|
34 |
+
"grad_norm": 0.878664493560791,
|
35 |
+
"learning_rate": 7.93034465929166e-06,
|
36 |
+
"loss": 1.5262,
|
37 |
+
"step": 40
|
38 |
+
},
|
39 |
+
{
|
40 |
+
"epoch": 0.5361930294906166,
|
41 |
+
"grad_norm": 0.6924058794975281,
|
42 |
+
"learning_rate": 7.848731934545733e-06,
|
43 |
+
"loss": 1.4716,
|
44 |
+
"step": 50
|
45 |
+
},
|
46 |
+
{
|
47 |
+
"epoch": 0.6434316353887399,
|
48 |
+
"grad_norm": 0.49218347668647766,
|
49 |
+
"learning_rate": 7.736655602515024e-06,
|
50 |
+
"loss": 1.3935,
|
51 |
+
"step": 60
|
52 |
+
},
|
53 |
+
{
|
54 |
+
"epoch": 0.7506702412868632,
|
55 |
+
"grad_norm": 0.48053157329559326,
|
56 |
+
"learning_rate": 7.5950027734017875e-06,
|
57 |
+
"loss": 1.3322,
|
58 |
+
"step": 70
|
59 |
+
},
|
60 |
+
{
|
61 |
+
"epoch": 0.8579088471849866,
|
62 |
+
"grad_norm": 0.4853557050228119,
|
63 |
+
"learning_rate": 7.424894662253031e-06,
|
64 |
+
"loss": 1.2899,
|
65 |
+
"step": 80
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.9651474530831099,
|
69 |
+
"grad_norm": 0.5471060276031494,
|
70 |
+
"learning_rate": 7.227677714282947e-06,
|
71 |
+
"loss": 1.284,
|
72 |
+
"step": 90
|
73 |
+
},
|
74 |
+
{
|
75 |
+
"epoch": 1.0723860589812333,
|
76 |
+
"grad_norm": 0.4984244406223297,
|
77 |
+
"learning_rate": 7.004912947446291e-06,
|
78 |
+
"loss": 1.2392,
|
79 |
+
"step": 100
|
80 |
+
},
|
81 |
+
{
|
82 |
+
"epoch": 1.1796246648793565,
|
83 |
+
"grad_norm": 0.5704559683799744,
|
84 |
+
"learning_rate": 6.758363596618656e-06,
|
85 |
+
"loss": 1.2033,
|
86 |
+
"step": 110
|
87 |
+
},
|
88 |
+
{
|
89 |
+
"epoch": 1.2868632707774799,
|
90 |
+
"grad_norm": 0.6068636178970337,
|
91 |
+
"learning_rate": 6.489981157182848e-06,
|
92 |
+
"loss": 1.1701,
|
93 |
+
"step": 120
|
94 |
+
},
|
95 |
+
{
|
96 |
+
"epoch": 1.3941018766756033,
|
97 |
+
"grad_norm": 0.5727308392524719,
|
98 |
+
"learning_rate": 6.2018899384896244e-06,
|
99 |
+
"loss": 1.1775,
|
100 |
+
"step": 130
|
101 |
+
},
|
102 |
+
{
|
103 |
+
"epoch": 1.5013404825737267,
|
104 |
+
"grad_norm": 0.542986273765564,
|
105 |
+
"learning_rate": 5.89637024945572e-06,
|
106 |
+
"loss": 1.1182,
|
107 |
+
"step": 140
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 1.6085790884718498,
|
111 |
+
"grad_norm": 0.5344908237457275,
|
112 |
+
"learning_rate": 5.575840349389077e-06,
|
113 |
+
"loss": 1.0901,
|
114 |
+
"step": 150
|
115 |
+
},
|
116 |
+
{
|
117 |
+
"epoch": 1.7158176943699732,
|
118 |
+
"grad_norm": 0.625162661075592,
|
119 |
+
"learning_rate": 5.242837306904692e-06,
|
120 |
+
"loss": 1.0817,
|
121 |
+
"step": 160
|
122 |
+
},
|
123 |
+
{
|
124 |
+
"epoch": 1.8230563002680964,
|
125 |
+
"grad_norm": 0.5978776812553406,
|
126 |
+
"learning_rate": 4.899996918437182e-06,
|
127 |
+
"loss": 1.0257,
|
128 |
+
"step": 170
|
129 |
+
},
|
130 |
+
{
|
131 |
+
"epoch": 1.9302949061662198,
|
132 |
+
"grad_norm": 0.6914421916007996,
|
133 |
+
"learning_rate": 4.550032845299744e-06,
|
134 |
+
"loss": 1.038,
|
135 |
+
"step": 180
|
136 |
+
},
|
137 |
+
{
|
138 |
+
"epoch": 2.037533512064343,
|
139 |
+
"grad_norm": 0.6677165627479553,
|
140 |
+
"learning_rate": 4.195715134424498e-06,
|
141 |
+
"loss": 1.012,
|
142 |
+
"step": 190
|
143 |
+
},
|
144 |
+
{
|
145 |
+
"epoch": 2.1447721179624666,
|
146 |
+
"grad_norm": 0.7082802057266235,
|
147 |
+
"learning_rate": 3.839848292797568e-06,
|
148 |
+
"loss": 0.9918,
|
149 |
+
"step": 200
|
150 |
+
},
|
151 |
+
{
|
152 |
+
"epoch": 2.1447721179624666,
|
153 |
+
"eval_loss": 0.9989345669746399,
|
154 |
+
"eval_runtime": 5.7346,
|
155 |
+
"eval_samples_per_second": 54.755,
|
156 |
+
"eval_steps_per_second": 1.744,
|
157 |
+
"step": 200
|
158 |
+
},
|
159 |
+
{
|
160 |
+
"epoch": 2.25201072386059,
|
161 |
+
"grad_norm": 0.6643340587615967,
|
162 |
+
"learning_rate": 3.485249089134823e-06,
|
163 |
+
"loss": 0.9689,
|
164 |
+
"step": 210
|
165 |
+
},
|
166 |
+
{
|
167 |
+
"epoch": 2.359249329758713,
|
168 |
+
"grad_norm": 0.6342740058898926,
|
169 |
+
"learning_rate": 3.134724258503188e-06,
|
170 |
+
"loss": 0.9704,
|
171 |
+
"step": 220
|
172 |
+
},
|
173 |
+
{
|
174 |
+
"epoch": 2.4664879356568363,
|
175 |
+
"grad_norm": 0.6604453921318054,
|
176 |
+
"learning_rate": 2.7910482863606053e-06,
|
177 |
+
"loss": 0.9555,
|
178 |
+
"step": 230
|
179 |
+
},
|
180 |
+
{
|
181 |
+
"epoch": 2.5737265415549597,
|
182 |
+
"grad_norm": 0.6870635747909546,
|
183 |
+
"learning_rate": 2.4569414478591614e-06,
|
184 |
+
"loss": 0.9478,
|
185 |
+
"step": 240
|
186 |
+
},
|
187 |
+
{
|
188 |
+
"epoch": 2.680965147453083,
|
189 |
+
"grad_norm": 0.6661774516105652,
|
190 |
+
"learning_rate": 2.1350482762353478e-06,
|
191 |
+
"loss": 0.9435,
|
192 |
+
"step": 250
|
193 |
+
},
|
194 |
+
{
|
195 |
+
"epoch": 2.7882037533512065,
|
196 |
+
"grad_norm": 0.7505815625190735,
|
197 |
+
"learning_rate": 1.827916630715212e-06,
|
198 |
+
"loss": 0.9339,
|
199 |
+
"step": 260
|
200 |
+
},
|
201 |
+
{
|
202 |
+
"epoch": 2.89544235924933,
|
203 |
+
"grad_norm": 0.6933086514472961,
|
204 |
+
"learning_rate": 1.5379775296167252e-06,
|
205 |
+
"loss": 0.9253,
|
206 |
+
"step": 270
|
207 |
+
},
|
208 |
+
{
|
209 |
+
"epoch": 3.002680965147453,
|
210 |
+
"grad_norm": 0.8025710582733154,
|
211 |
+
"learning_rate": 1.2675259082751067e-06,
|
212 |
+
"loss": 0.9055,
|
213 |
+
"step": 280
|
214 |
+
},
|
215 |
+
{
|
216 |
+
"epoch": 3.1099195710455763,
|
217 |
+
"grad_norm": 0.7908952236175537,
|
218 |
+
"learning_rate": 1.018702454096546e-06,
|
219 |
+
"loss": 0.9039,
|
220 |
+
"step": 290
|
221 |
+
},
|
222 |
+
{
|
223 |
+
"epoch": 3.2171581769436997,
|
224 |
+
"grad_norm": 0.7214229702949524,
|
225 |
+
"learning_rate": 7.934766625201255e-07,
|
226 |
+
"loss": 0.9274,
|
227 |
+
"step": 300
|
228 |
+
},
|
229 |
+
{
|
230 |
+
"epoch": 3.324396782841823,
|
231 |
+
"grad_norm": 0.7900909781455994,
|
232 |
+
"learning_rate": 5.936312480039261e-07,
|
233 |
+
"loss": 0.905,
|
234 |
+
"step": 310
|
235 |
+
},
|
236 |
+
{
|
237 |
+
"epoch": 3.4316353887399464,
|
238 |
+
"grad_norm": 0.6883360147476196,
|
239 |
+
"learning_rate": 4.207480334259821e-07,
|
240 |
+
"loss": 0.9131,
|
241 |
+
"step": 320
|
242 |
+
},
|
243 |
+
{
|
244 |
+
"epoch": 3.53887399463807,
|
245 |
+
"grad_norm": 0.7782545685768127,
|
246 |
+
"learning_rate": 2.761954295887592e-07,
|
247 |
+
"loss": 0.9053,
|
248 |
+
"step": 330
|
249 |
+
},
|
250 |
+
{
|
251 |
+
"epoch": 3.646112600536193,
|
252 |
+
"grad_norm": 0.7688133716583252,
|
253 |
+
"learning_rate": 1.6111760392979012e-07,
|
254 |
+
"loss": 0.8879,
|
255 |
+
"step": 340
|
256 |
+
},
|
257 |
+
{
|
258 |
+
"epoch": 3.753351206434316,
|
259 |
+
"grad_norm": 0.8082988262176514,
|
260 |
+
"learning_rate": 7.642542417061859e-08,
|
261 |
+
"loss": 0.9101,
|
262 |
+
"step": 350
|
263 |
+
},
|
264 |
+
{
|
265 |
+
"epoch": 3.8605898123324396,
|
266 |
+
"grad_norm": 0.7244554162025452,
|
267 |
+
"learning_rate": 2.2789248587207655e-08,
|
268 |
+
"loss": 0.8842,
|
269 |
+
"step": 360
|
270 |
+
},
|
271 |
+
{
|
272 |
+
"epoch": 3.967828418230563,
|
273 |
+
"grad_norm": 0.8839091062545776,
|
274 |
+
"learning_rate": 6.33619968457566e-10,
|
275 |
+
"loss": 0.9048,
|
276 |
+
"step": 370
|
277 |
+
},
|
278 |
+
{
|
279 |
+
"epoch": 3.9892761394101877,
|
280 |
+
"step": 372,
|
281 |
+
"total_flos": 3.7910142911093146e+17,
|
282 |
+
"train_loss": 1.1117366898444392,
|
283 |
+
"train_runtime": 1080.8388,
|
284 |
+
"train_samples_per_second": 22.064,
|
285 |
+
"train_steps_per_second": 0.344
|
286 |
+
}
|
287 |
+
],
|
288 |
+
"logging_steps": 10,
|
289 |
+
"max_steps": 372,
|
290 |
+
"num_input_tokens_seen": 0,
|
291 |
+
"num_train_epochs": 4,
|
292 |
+
"save_steps": 200,
|
293 |
+
"stateful_callbacks": {
|
294 |
+
"TrainerControl": {
|
295 |
+
"args": {
|
296 |
+
"should_epoch_stop": false,
|
297 |
+
"should_evaluate": false,
|
298 |
+
"should_log": false,
|
299 |
+
"should_save": true,
|
300 |
+
"should_training_stop": true
|
301 |
+
},
|
302 |
+
"attributes": {}
|
303 |
+
}
|
304 |
+
},
|
305 |
+
"total_flos": 3.7910142911093146e+17,
|
306 |
+
"train_batch_size": 4,
|
307 |
+
"trial_name": null,
|
308 |
+
"trial_params": null
|
309 |
+
}
|
training_eval_loss.png
ADDED
training_loss.png
ADDED