End of training
Browse files- README.md +1 -1
- all_results.json +17 -0
- eval_results.json +11 -0
- train_results.json +10 -0
- trainer_state.json +297 -0
README.md
CHANGED
@@ -16,7 +16,7 @@ should probably proofread and complete it, then remove this comment. -->
|
|
16 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/pszemraj/long-generation-tests/runs/ethp25f9)
|
17 |
# Qwen2-1.5B-stepbasin-books-vN
|
18 |
|
19 |
-
This model is a fine-tuned version of [Qwen/Qwen2-1.5B](https://huggingface.co/Qwen/Qwen2-1.5B) on
|
20 |
It achieves the following results on the evaluation set:
|
21 |
- Loss: 2.8110
|
22 |
- Accuracy: 0.4298
|
|
|
16 |
[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="200" height="32"/>](https://wandb.ai/pszemraj/long-generation-tests/runs/ethp25f9)
|
17 |
# Qwen2-1.5B-stepbasin-books-vN
|
18 |
|
19 |
+
This model is a fine-tuned version of [Qwen/Qwen2-1.5B](https://huggingface.co/Qwen/Qwen2-1.5B) on the BEE-spoke-data/stepbasin-books dataset.
|
20 |
It achieves the following results on the evaluation set:
|
21 |
- Loss: 2.8110
|
22 |
- Accuracy: 0.4298
|
all_results.json
ADDED
@@ -0,0 +1,17 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 2.9899888765294773,
|
3 |
+
"eval_accuracy": 0.4297979192055684,
|
4 |
+
"eval_loss": 2.810983180999756,
|
5 |
+
"eval_runtime": 19.3217,
|
6 |
+
"eval_samples": 29,
|
7 |
+
"eval_samples_per_second": 1.501,
|
8 |
+
"eval_steps_per_second": 1.501,
|
9 |
+
"num_input_tokens_seen": 44040192,
|
10 |
+
"perplexity": 16.62625680536472,
|
11 |
+
"total_flos": 3.462459117703004e+17,
|
12 |
+
"train_loss": 2.7097946518943425,
|
13 |
+
"train_runtime": 6795.1554,
|
14 |
+
"train_samples": 899,
|
15 |
+
"train_samples_per_second": 0.397,
|
16 |
+
"train_steps_per_second": 0.012
|
17 |
+
}
|
eval_results.json
ADDED
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 2.9899888765294773,
|
3 |
+
"eval_accuracy": 0.4297979192055684,
|
4 |
+
"eval_loss": 2.810983180999756,
|
5 |
+
"eval_runtime": 19.3217,
|
6 |
+
"eval_samples": 29,
|
7 |
+
"eval_samples_per_second": 1.501,
|
8 |
+
"eval_steps_per_second": 1.501,
|
9 |
+
"num_input_tokens_seen": 44040192,
|
10 |
+
"perplexity": 16.62625680536472
|
11 |
+
}
|
train_results.json
ADDED
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"epoch": 2.9899888765294773,
|
3 |
+
"num_input_tokens_seen": 44040192,
|
4 |
+
"total_flos": 3.462459117703004e+17,
|
5 |
+
"train_loss": 2.7097946518943425,
|
6 |
+
"train_runtime": 6795.1554,
|
7 |
+
"train_samples": 899,
|
8 |
+
"train_samples_per_second": 0.397,
|
9 |
+
"train_steps_per_second": 0.012
|
10 |
+
}
|
trainer_state.json
ADDED
@@ -0,0 +1,297 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
{
|
2 |
+
"best_metric": null,
|
3 |
+
"best_model_checkpoint": null,
|
4 |
+
"epoch": 2.9899888765294773,
|
5 |
+
"eval_steps": 300.0,
|
6 |
+
"global_step": 84,
|
7 |
+
"is_hyper_param_search": false,
|
8 |
+
"is_local_process_zero": true,
|
9 |
+
"is_world_process_zero": true,
|
10 |
+
"log_history": [
|
11 |
+
{
|
12 |
+
"epoch": 0.1067853170189099,
|
13 |
+
"grad_norm": 0.5859375,
|
14 |
+
"learning_rate": 1.8e-05,
|
15 |
+
"loss": 2.8121,
|
16 |
+
"num_input_tokens_seen": 1572864,
|
17 |
+
"step": 3
|
18 |
+
},
|
19 |
+
{
|
20 |
+
"epoch": 0.2135706340378198,
|
21 |
+
"grad_norm": 0.69921875,
|
22 |
+
"learning_rate": 2.9988140958409528e-05,
|
23 |
+
"loss": 2.8211,
|
24 |
+
"num_input_tokens_seen": 3145728,
|
25 |
+
"step": 6
|
26 |
+
},
|
27 |
+
{
|
28 |
+
"epoch": 0.3203559510567297,
|
29 |
+
"grad_norm": 0.796875,
|
30 |
+
"learning_rate": 2.9810630129045003e-05,
|
31 |
+
"loss": 2.7286,
|
32 |
+
"num_input_tokens_seen": 4718592,
|
33 |
+
"step": 9
|
34 |
+
},
|
35 |
+
{
|
36 |
+
"epoch": 0.4271412680756396,
|
37 |
+
"grad_norm": 0.71875,
|
38 |
+
"learning_rate": 2.9422573564911305e-05,
|
39 |
+
"loss": 2.7688,
|
40 |
+
"num_input_tokens_seen": 6291456,
|
41 |
+
"step": 12
|
42 |
+
},
|
43 |
+
{
|
44 |
+
"epoch": 0.5339265850945495,
|
45 |
+
"grad_norm": 0.58984375,
|
46 |
+
"learning_rate": 2.88294878209231e-05,
|
47 |
+
"loss": 2.756,
|
48 |
+
"num_input_tokens_seen": 7864320,
|
49 |
+
"step": 15
|
50 |
+
},
|
51 |
+
{
|
52 |
+
"epoch": 0.6407119021134594,
|
53 |
+
"grad_norm": 0.54296875,
|
54 |
+
"learning_rate": 2.8039804116593743e-05,
|
55 |
+
"loss": 2.7252,
|
56 |
+
"num_input_tokens_seen": 9437184,
|
57 |
+
"step": 18
|
58 |
+
},
|
59 |
+
{
|
60 |
+
"epoch": 0.7474972191323693,
|
61 |
+
"grad_norm": 0.54296875,
|
62 |
+
"learning_rate": 2.7064748479061476e-05,
|
63 |
+
"loss": 2.7188,
|
64 |
+
"num_input_tokens_seen": 11010048,
|
65 |
+
"step": 21
|
66 |
+
},
|
67 |
+
{
|
68 |
+
"epoch": 0.8542825361512792,
|
69 |
+
"grad_norm": 0.484375,
|
70 |
+
"learning_rate": 2.5918182155542415e-05,
|
71 |
+
"loss": 2.7574,
|
72 |
+
"num_input_tokens_seen": 12582912,
|
73 |
+
"step": 24
|
74 |
+
},
|
75 |
+
{
|
76 |
+
"epoch": 0.9610678531701891,
|
77 |
+
"grad_norm": 0.5,
|
78 |
+
"learning_rate": 2.4616404563883302e-05,
|
79 |
+
"loss": 2.7792,
|
80 |
+
"num_input_tokens_seen": 14155776,
|
81 |
+
"step": 27
|
82 |
+
},
|
83 |
+
{
|
84 |
+
"epoch": 0.996662958843159,
|
85 |
+
"eval_accuracy": 0.4287265815910942,
|
86 |
+
"eval_loss": 2.818293809890747,
|
87 |
+
"eval_runtime": 19.7558,
|
88 |
+
"eval_samples_per_second": 1.468,
|
89 |
+
"eval_steps_per_second": 1.468,
|
90 |
+
"num_input_tokens_seen": 14729216,
|
91 |
+
"step": 28
|
92 |
+
},
|
93 |
+
{
|
94 |
+
"epoch": 1.067853170189099,
|
95 |
+
"grad_norm": 0.50390625,
|
96 |
+
"learning_rate": 2.3177921582440015e-05,
|
97 |
+
"loss": 2.6953,
|
98 |
+
"num_input_tokens_seen": 15728640,
|
99 |
+
"step": 30
|
100 |
+
},
|
101 |
+
{
|
102 |
+
"epoch": 1.174638487208009,
|
103 |
+
"grad_norm": 0.462890625,
|
104 |
+
"learning_rate": 2.162318247323868e-05,
|
105 |
+
"loss": 2.7098,
|
106 |
+
"num_input_tokens_seen": 17301504,
|
107 |
+
"step": 33
|
108 |
+
},
|
109 |
+
{
|
110 |
+
"epoch": 1.2814238042269188,
|
111 |
+
"grad_norm": 0.451171875,
|
112 |
+
"learning_rate": 1.997428917828102e-05,
|
113 |
+
"loss": 2.6726,
|
114 |
+
"num_input_tokens_seen": 18874368,
|
115 |
+
"step": 36
|
116 |
+
},
|
117 |
+
{
|
118 |
+
"epoch": 1.3882091212458287,
|
119 |
+
"grad_norm": 0.462890625,
|
120 |
+
"learning_rate": 1.825468212159477e-05,
|
121 |
+
"loss": 2.6655,
|
122 |
+
"num_input_tokens_seen": 20447232,
|
123 |
+
"step": 39
|
124 |
+
},
|
125 |
+
{
|
126 |
+
"epoch": 1.4949944382647387,
|
127 |
+
"grad_norm": 0.455078125,
|
128 |
+
"learning_rate": 1.6488806983620927e-05,
|
129 |
+
"loss": 2.7292,
|
130 |
+
"num_input_tokens_seen": 22020096,
|
131 |
+
"step": 42
|
132 |
+
},
|
133 |
+
{
|
134 |
+
"epoch": 1.6017797552836486,
|
135 |
+
"grad_norm": 0.4453125,
|
136 |
+
"learning_rate": 1.4701767185023948e-05,
|
137 |
+
"loss": 2.6312,
|
138 |
+
"num_input_tokens_seen": 23592960,
|
139 |
+
"step": 45
|
140 |
+
},
|
141 |
+
{
|
142 |
+
"epoch": 1.7085650723025583,
|
143 |
+
"grad_norm": 0.43359375,
|
144 |
+
"learning_rate": 1.2918967020163978e-05,
|
145 |
+
"loss": 2.7176,
|
146 |
+
"num_input_tokens_seen": 25165824,
|
147 |
+
"step": 48
|
148 |
+
},
|
149 |
+
{
|
150 |
+
"epoch": 1.8153503893214684,
|
151 |
+
"grad_norm": 0.4453125,
|
152 |
+
"learning_rate": 1.116575051339288e-05,
|
153 |
+
"loss": 2.7234,
|
154 |
+
"num_input_tokens_seen": 26738688,
|
155 |
+
"step": 51
|
156 |
+
},
|
157 |
+
{
|
158 |
+
"epoch": 1.9221357063403781,
|
159 |
+
"grad_norm": 0.4375,
|
160 |
+
"learning_rate": 9.467041132139884e-06,
|
161 |
+
"loss": 2.6971,
|
162 |
+
"num_input_tokens_seen": 28311552,
|
163 |
+
"step": 54
|
164 |
+
},
|
165 |
+
{
|
166 |
+
"epoch": 1.993325917686318,
|
167 |
+
"eval_accuracy": 0.4297116228554831,
|
168 |
+
"eval_loss": 2.8111748695373535,
|
169 |
+
"eval_runtime": 19.5837,
|
170 |
+
"eval_samples_per_second": 1.481,
|
171 |
+
"eval_steps_per_second": 1.481,
|
172 |
+
"num_input_tokens_seen": 29458432,
|
173 |
+
"step": 56
|
174 |
+
},
|
175 |
+
{
|
176 |
+
"epoch": 2.0289210233592883,
|
177 |
+
"grad_norm": 0.41796875,
|
178 |
+
"learning_rate": 7.846987478572411e-06,
|
179 |
+
"loss": 2.6623,
|
180 |
+
"num_input_tokens_seen": 29884416,
|
181 |
+
"step": 57
|
182 |
+
},
|
183 |
+
{
|
184 |
+
"epoch": 2.135706340378198,
|
185 |
+
"grad_norm": 0.396484375,
|
186 |
+
"learning_rate": 6.328619996627272e-06,
|
187 |
+
"loss": 2.6211,
|
188 |
+
"num_input_tokens_seen": 31457280,
|
189 |
+
"step": 60
|
190 |
+
},
|
191 |
+
{
|
192 |
+
"epoch": 2.242491657397108,
|
193 |
+
"grad_norm": 0.416015625,
|
194 |
+
"learning_rate": 4.933523574614447e-06,
|
195 |
+
"loss": 2.6728,
|
196 |
+
"num_input_tokens_seen": 33030144,
|
197 |
+
"step": 63
|
198 |
+
},
|
199 |
+
{
|
200 |
+
"epoch": 2.349276974416018,
|
201 |
+
"grad_norm": 0.41015625,
|
202 |
+
"learning_rate": 3.6815306976265466e-06,
|
203 |
+
"loss": 2.6567,
|
204 |
+
"num_input_tokens_seen": 34603008,
|
205 |
+
"step": 66
|
206 |
+
},
|
207 |
+
{
|
208 |
+
"epoch": 2.456062291434928,
|
209 |
+
"grad_norm": 0.435546875,
|
210 |
+
"learning_rate": 2.590439511854144e-06,
|
211 |
+
"loss": 2.6862,
|
212 |
+
"num_input_tokens_seen": 36175872,
|
213 |
+
"step": 69
|
214 |
+
},
|
215 |
+
{
|
216 |
+
"epoch": 2.5628476084538376,
|
217 |
+
"grad_norm": 0.419921875,
|
218 |
+
"learning_rate": 1.6757608087630249e-06,
|
219 |
+
"loss": 2.6841,
|
220 |
+
"num_input_tokens_seen": 37748736,
|
221 |
+
"step": 72
|
222 |
+
},
|
223 |
+
{
|
224 |
+
"epoch": 2.6696329254727473,
|
225 |
+
"grad_norm": 0.40625,
|
226 |
+
"learning_rate": 9.504975259690835e-07,
|
227 |
+
"loss": 2.7272,
|
228 |
+
"num_input_tokens_seen": 39321600,
|
229 |
+
"step": 75
|
230 |
+
},
|
231 |
+
{
|
232 |
+
"epoch": 2.7764182424916575,
|
233 |
+
"grad_norm": 0.40625,
|
234 |
+
"learning_rate": 4.2495989939384916e-07,
|
235 |
+
"loss": 2.6643,
|
236 |
+
"num_input_tokens_seen": 40894464,
|
237 |
+
"step": 78
|
238 |
+
},
|
239 |
+
{
|
240 |
+
"epoch": 2.883203559510567,
|
241 |
+
"grad_norm": 0.408203125,
|
242 |
+
"learning_rate": 1.0661889447039886e-07,
|
243 |
+
"loss": 2.679,
|
244 |
+
"num_input_tokens_seen": 42467328,
|
245 |
+
"step": 81
|
246 |
+
},
|
247 |
+
{
|
248 |
+
"epoch": 2.9899888765294773,
|
249 |
+
"grad_norm": 0.412109375,
|
250 |
+
"learning_rate": 0.0,
|
251 |
+
"loss": 2.7116,
|
252 |
+
"num_input_tokens_seen": 44040192,
|
253 |
+
"step": 84
|
254 |
+
},
|
255 |
+
{
|
256 |
+
"epoch": 2.9899888765294773,
|
257 |
+
"eval_accuracy": 0.4297979192055684,
|
258 |
+
"eval_loss": 2.810983180999756,
|
259 |
+
"eval_runtime": 19.4794,
|
260 |
+
"eval_samples_per_second": 1.489,
|
261 |
+
"eval_steps_per_second": 1.489,
|
262 |
+
"num_input_tokens_seen": 44040192,
|
263 |
+
"step": 84
|
264 |
+
},
|
265 |
+
{
|
266 |
+
"epoch": 2.9899888765294773,
|
267 |
+
"num_input_tokens_seen": 44040192,
|
268 |
+
"step": 84,
|
269 |
+
"total_flos": 3.462459117703004e+17,
|
270 |
+
"train_loss": 2.7097946518943425,
|
271 |
+
"train_runtime": 6795.1554,
|
272 |
+
"train_samples_per_second": 0.397,
|
273 |
+
"train_steps_per_second": 0.012
|
274 |
+
}
|
275 |
+
],
|
276 |
+
"logging_steps": 3,
|
277 |
+
"max_steps": 84,
|
278 |
+
"num_input_tokens_seen": 44040192,
|
279 |
+
"num_train_epochs": 3,
|
280 |
+
"save_steps": 50,
|
281 |
+
"stateful_callbacks": {
|
282 |
+
"TrainerControl": {
|
283 |
+
"args": {
|
284 |
+
"should_epoch_stop": false,
|
285 |
+
"should_evaluate": false,
|
286 |
+
"should_log": false,
|
287 |
+
"should_save": true,
|
288 |
+
"should_training_stop": true
|
289 |
+
},
|
290 |
+
"attributes": {}
|
291 |
+
}
|
292 |
+
},
|
293 |
+
"total_flos": 3.462459117703004e+17,
|
294 |
+
"train_batch_size": 1,
|
295 |
+
"trial_name": null,
|
296 |
+
"trial_params": null
|
297 |
+
}
|