Draup-DS commited on
Commit
3b56547
·
verified ·
1 Parent(s): 5355c40

Upload folder using huggingface_hub

Browse files
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer.json filter=lfs diff=lfs merge=lfs -text
config.json ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_sliding_window_pattern": 6,
3
+ "architectures": [
4
+ "Gemma3ForCausalLM"
5
+ ],
6
+ "attention_bias": false,
7
+ "attention_dropout": 0.0,
8
+ "attn_logit_softcapping": null,
9
+ "bos_token_id": 2,
10
+ "dtype": "float32",
11
+ "eos_token_id": 1,
12
+ "final_logit_softcapping": null,
13
+ "head_dim": 256,
14
+ "hidden_activation": "gelu_pytorch_tanh",
15
+ "hidden_size": 640,
16
+ "initializer_range": 0.02,
17
+ "intermediate_size": 2048,
18
+ "layer_types": [
19
+ "sliding_attention",
20
+ "sliding_attention",
21
+ "sliding_attention",
22
+ "sliding_attention",
23
+ "sliding_attention",
24
+ "full_attention",
25
+ "sliding_attention",
26
+ "sliding_attention",
27
+ "sliding_attention",
28
+ "sliding_attention",
29
+ "sliding_attention",
30
+ "full_attention",
31
+ "sliding_attention",
32
+ "sliding_attention",
33
+ "sliding_attention",
34
+ "sliding_attention",
35
+ "sliding_attention",
36
+ "full_attention"
37
+ ],
38
+ "max_position_embeddings": 32768,
39
+ "model_type": "gemma3_text",
40
+ "num_attention_heads": 4,
41
+ "num_hidden_layers": 18,
42
+ "num_key_value_heads": 1,
43
+ "pad_token_id": 0,
44
+ "query_pre_attn_scalar": 256,
45
+ "rms_norm_eps": 1e-06,
46
+ "rope_local_base_freq": 10000.0,
47
+ "rope_scaling": null,
48
+ "rope_theta": 1000000.0,
49
+ "sliding_window": 512,
50
+ "transformers_version": "4.56.2",
51
+ "use_bidirectional_attention": false,
52
+ "use_cache": true,
53
+ "vocab_size": 262144
54
+ }
generation_config.json ADDED
@@ -0,0 +1,12 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "bos_token_id": 2,
3
+ "cache_implementation": "hybrid",
4
+ "do_sample": true,
5
+ "eos_token_id": [
6
+ 1
7
+ ],
8
+ "pad_token_id": 0,
9
+ "top_k": 64,
10
+ "top_p": 0.95,
11
+ "transformers_version": "4.56.2"
12
+ }
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:62a73afe2ffa8e71f07d1c152f0d57b49d4bcd88df83230d2dbd5bcb07a6c79d
3
+ size 1072419256
optimizer.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:35c9c0c617a817d1cf4fa9ab35b38c29714198790133142480bd1c18f813cb30
3
+ size 2144983354
rng_state.pth ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4ab4163c79d02a2b1cd5e87832e03656d66b051535f06b7e0f0faade52cd7bd7
3
+ size 14244
scheduler.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d966b43422964246a7405d68c8d611f446291fba9748aa50ad3293aec861b68d
3
+ size 1064
special_tokens_map.json ADDED
@@ -0,0 +1,33 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "boi_token": "<start_of_image>",
3
+ "bos_token": {
4
+ "content": "<bos>",
5
+ "lstrip": false,
6
+ "normalized": false,
7
+ "rstrip": false,
8
+ "single_word": false
9
+ },
10
+ "eoi_token": "<end_of_image>",
11
+ "eos_token": {
12
+ "content": "<eos>",
13
+ "lstrip": false,
14
+ "normalized": false,
15
+ "rstrip": false,
16
+ "single_word": false
17
+ },
18
+ "image_token": "<image_soft_token>",
19
+ "pad_token": {
20
+ "content": "<pad>",
21
+ "lstrip": false,
22
+ "normalized": false,
23
+ "rstrip": false,
24
+ "single_word": false
25
+ },
26
+ "unk_token": {
27
+ "content": "<unk>",
28
+ "lstrip": false,
29
+ "normalized": false,
30
+ "rstrip": false,
31
+ "single_word": false
32
+ }
33
+ }
tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e26690f066f0d0683dab7b240a534612f1d93dacee419e19dfe7e296548ec878
3
+ size 33384666
tokenizer_config.json ADDED
The diff for this file is too large to render. See raw diff
 
trainer_state.json ADDED
@@ -0,0 +1,899 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "best_global_step": null,
3
+ "best_metric": null,
4
+ "best_model_checkpoint": null,
5
+ "epoch": 4.0,
6
+ "eval_steps": 500,
7
+ "global_step": 11900,
8
+ "is_hyper_param_search": false,
9
+ "is_local_process_zero": true,
10
+ "is_world_process_zero": true,
11
+ "log_history": [
12
+ {
13
+ "epoch": 0.03361344537815126,
14
+ "grad_norm": 1.2532932758331299,
15
+ "learning_rate": 4.966722689075631e-05,
16
+ "loss": 0.7069,
17
+ "step": 100
18
+ },
19
+ {
20
+ "epoch": 0.06722689075630252,
21
+ "grad_norm": 0.3335288465023041,
22
+ "learning_rate": 4.933109243697479e-05,
23
+ "loss": 0.0112,
24
+ "step": 200
25
+ },
26
+ {
27
+ "epoch": 0.10084033613445378,
28
+ "grad_norm": 0.21957238018512726,
29
+ "learning_rate": 4.899495798319328e-05,
30
+ "loss": 0.0075,
31
+ "step": 300
32
+ },
33
+ {
34
+ "epoch": 0.13445378151260504,
35
+ "grad_norm": 0.12609940767288208,
36
+ "learning_rate": 4.8658823529411766e-05,
37
+ "loss": 0.0063,
38
+ "step": 400
39
+ },
40
+ {
41
+ "epoch": 0.16806722689075632,
42
+ "grad_norm": 0.25657421350479126,
43
+ "learning_rate": 4.832268907563026e-05,
44
+ "loss": 0.0056,
45
+ "step": 500
46
+ },
47
+ {
48
+ "epoch": 0.20168067226890757,
49
+ "grad_norm": 0.5225902795791626,
50
+ "learning_rate": 4.7986554621848737e-05,
51
+ "loss": 0.0073,
52
+ "step": 600
53
+ },
54
+ {
55
+ "epoch": 0.23529411764705882,
56
+ "grad_norm": 0.3552115261554718,
57
+ "learning_rate": 4.765042016806723e-05,
58
+ "loss": 0.0049,
59
+ "step": 700
60
+ },
61
+ {
62
+ "epoch": 0.2689075630252101,
63
+ "grad_norm": 0.4207090139389038,
64
+ "learning_rate": 4.7314285714285714e-05,
65
+ "loss": 0.0043,
66
+ "step": 800
67
+ },
68
+ {
69
+ "epoch": 0.3025210084033613,
70
+ "grad_norm": 0.6347660422325134,
71
+ "learning_rate": 4.6978151260504207e-05,
72
+ "loss": 0.0039,
73
+ "step": 900
74
+ },
75
+ {
76
+ "epoch": 0.33613445378151263,
77
+ "grad_norm": 0.4249831438064575,
78
+ "learning_rate": 4.664201680672269e-05,
79
+ "loss": 0.0035,
80
+ "step": 1000
81
+ },
82
+ {
83
+ "epoch": 0.3697478991596639,
84
+ "grad_norm": 0.20205572247505188,
85
+ "learning_rate": 4.630588235294118e-05,
86
+ "loss": 0.0045,
87
+ "step": 1100
88
+ },
89
+ {
90
+ "epoch": 0.40336134453781514,
91
+ "grad_norm": 0.25402364134788513,
92
+ "learning_rate": 4.596974789915966e-05,
93
+ "loss": 0.0048,
94
+ "step": 1200
95
+ },
96
+ {
97
+ "epoch": 0.4369747899159664,
98
+ "grad_norm": 0.11454110592603683,
99
+ "learning_rate": 4.5633613445378155e-05,
100
+ "loss": 0.0046,
101
+ "step": 1300
102
+ },
103
+ {
104
+ "epoch": 0.47058823529411764,
105
+ "grad_norm": 0.29738855361938477,
106
+ "learning_rate": 4.529747899159664e-05,
107
+ "loss": 0.0084,
108
+ "step": 1400
109
+ },
110
+ {
111
+ "epoch": 0.5042016806722689,
112
+ "grad_norm": 0.048551127314567566,
113
+ "learning_rate": 4.4961344537815126e-05,
114
+ "loss": 0.0029,
115
+ "step": 1500
116
+ },
117
+ {
118
+ "epoch": 0.5378151260504201,
119
+ "grad_norm": 0.1416209638118744,
120
+ "learning_rate": 4.462521008403361e-05,
121
+ "loss": 0.0036,
122
+ "step": 1600
123
+ },
124
+ {
125
+ "epoch": 0.5714285714285714,
126
+ "grad_norm": 0.4617021083831787,
127
+ "learning_rate": 4.4289075630252104e-05,
128
+ "loss": 0.0029,
129
+ "step": 1700
130
+ },
131
+ {
132
+ "epoch": 0.6050420168067226,
133
+ "grad_norm": 0.25475579500198364,
134
+ "learning_rate": 4.395294117647059e-05,
135
+ "loss": 0.0031,
136
+ "step": 1800
137
+ },
138
+ {
139
+ "epoch": 0.6386554621848739,
140
+ "grad_norm": 0.23906832933425903,
141
+ "learning_rate": 4.361680672268908e-05,
142
+ "loss": 0.0032,
143
+ "step": 1900
144
+ },
145
+ {
146
+ "epoch": 0.6722689075630253,
147
+ "grad_norm": 0.2611936032772064,
148
+ "learning_rate": 4.328067226890756e-05,
149
+ "loss": 0.0037,
150
+ "step": 2000
151
+ },
152
+ {
153
+ "epoch": 0.7058823529411765,
154
+ "grad_norm": 0.04429571330547333,
155
+ "learning_rate": 4.294453781512605e-05,
156
+ "loss": 0.0028,
157
+ "step": 2100
158
+ },
159
+ {
160
+ "epoch": 0.7394957983193278,
161
+ "grad_norm": 0.47502580285072327,
162
+ "learning_rate": 4.260840336134454e-05,
163
+ "loss": 0.0032,
164
+ "step": 2200
165
+ },
166
+ {
167
+ "epoch": 0.773109243697479,
168
+ "grad_norm": 0.2212914377450943,
169
+ "learning_rate": 4.227226890756303e-05,
170
+ "loss": 0.0024,
171
+ "step": 2300
172
+ },
173
+ {
174
+ "epoch": 0.8067226890756303,
175
+ "grad_norm": 20.17454719543457,
176
+ "learning_rate": 4.1936134453781516e-05,
177
+ "loss": 0.0057,
178
+ "step": 2400
179
+ },
180
+ {
181
+ "epoch": 0.8403361344537815,
182
+ "grad_norm": 0.48056426644325256,
183
+ "learning_rate": 4.16e-05,
184
+ "loss": 0.0039,
185
+ "step": 2500
186
+ },
187
+ {
188
+ "epoch": 0.8739495798319328,
189
+ "grad_norm": 0.4135717451572418,
190
+ "learning_rate": 4.126386554621849e-05,
191
+ "loss": 0.0029,
192
+ "step": 2600
193
+ },
194
+ {
195
+ "epoch": 0.907563025210084,
196
+ "grad_norm": 0.10312103480100632,
197
+ "learning_rate": 4.092773109243698e-05,
198
+ "loss": 0.0027,
199
+ "step": 2700
200
+ },
201
+ {
202
+ "epoch": 0.9411764705882353,
203
+ "grad_norm": 0.10092929005622864,
204
+ "learning_rate": 4.0591596638655465e-05,
205
+ "loss": 0.003,
206
+ "step": 2800
207
+ },
208
+ {
209
+ "epoch": 0.9747899159663865,
210
+ "grad_norm": 0.42715680599212646,
211
+ "learning_rate": 4.025546218487395e-05,
212
+ "loss": 0.0036,
213
+ "step": 2900
214
+ },
215
+ {
216
+ "epoch": 1.0,
217
+ "eval_loss": 0.0022414932027459145,
218
+ "eval_runtime": 62.59,
219
+ "eval_samples_per_second": 124.908,
220
+ "eval_steps_per_second": 7.813,
221
+ "step": 2975
222
+ },
223
+ {
224
+ "epoch": 1.0084033613445378,
225
+ "grad_norm": 0.010128838941454887,
226
+ "learning_rate": 3.9919327731092436e-05,
227
+ "loss": 0.0025,
228
+ "step": 3000
229
+ },
230
+ {
231
+ "epoch": 1.0420168067226891,
232
+ "grad_norm": 0.05345134809613228,
233
+ "learning_rate": 3.958319327731093e-05,
234
+ "loss": 0.0014,
235
+ "step": 3100
236
+ },
237
+ {
238
+ "epoch": 1.0756302521008403,
239
+ "grad_norm": 0.030169779434800148,
240
+ "learning_rate": 3.9247058823529413e-05,
241
+ "loss": 0.0019,
242
+ "step": 3200
243
+ },
244
+ {
245
+ "epoch": 1.1092436974789917,
246
+ "grad_norm": 0.26832690834999084,
247
+ "learning_rate": 3.8910924369747906e-05,
248
+ "loss": 0.0017,
249
+ "step": 3300
250
+ },
251
+ {
252
+ "epoch": 1.1428571428571428,
253
+ "grad_norm": 0.0775144025683403,
254
+ "learning_rate": 3.8574789915966384e-05,
255
+ "loss": 0.002,
256
+ "step": 3400
257
+ },
258
+ {
259
+ "epoch": 1.1764705882352942,
260
+ "grad_norm": 0.07601425796747208,
261
+ "learning_rate": 3.823865546218488e-05,
262
+ "loss": 0.0017,
263
+ "step": 3500
264
+ },
265
+ {
266
+ "epoch": 1.2100840336134453,
267
+ "grad_norm": 0.23803076148033142,
268
+ "learning_rate": 3.790252100840336e-05,
269
+ "loss": 0.0021,
270
+ "step": 3600
271
+ },
272
+ {
273
+ "epoch": 1.2436974789915967,
274
+ "grad_norm": 0.014035782776772976,
275
+ "learning_rate": 3.7566386554621854e-05,
276
+ "loss": 0.0019,
277
+ "step": 3700
278
+ },
279
+ {
280
+ "epoch": 1.2773109243697478,
281
+ "grad_norm": 0.15406791865825653,
282
+ "learning_rate": 3.723025210084034e-05,
283
+ "loss": 0.002,
284
+ "step": 3800
285
+ },
286
+ {
287
+ "epoch": 1.3109243697478992,
288
+ "grad_norm": 0.014077406376600266,
289
+ "learning_rate": 3.6894117647058825e-05,
290
+ "loss": 0.0019,
291
+ "step": 3900
292
+ },
293
+ {
294
+ "epoch": 1.3445378151260505,
295
+ "grad_norm": 0.06503783166408539,
296
+ "learning_rate": 3.655798319327731e-05,
297
+ "loss": 0.0014,
298
+ "step": 4000
299
+ },
300
+ {
301
+ "epoch": 1.3781512605042017,
302
+ "grad_norm": 0.0875903069972992,
303
+ "learning_rate": 3.62218487394958e-05,
304
+ "loss": 0.0019,
305
+ "step": 4100
306
+ },
307
+ {
308
+ "epoch": 1.4117647058823528,
309
+ "grad_norm": 0.010527476668357849,
310
+ "learning_rate": 3.588571428571429e-05,
311
+ "loss": 0.0017,
312
+ "step": 4200
313
+ },
314
+ {
315
+ "epoch": 1.4453781512605042,
316
+ "grad_norm": 0.011578446254134178,
317
+ "learning_rate": 3.5549579831932774e-05,
318
+ "loss": 0.0024,
319
+ "step": 4300
320
+ },
321
+ {
322
+ "epoch": 1.4789915966386555,
323
+ "grad_norm": 0.0051996163092553616,
324
+ "learning_rate": 3.521344537815126e-05,
325
+ "loss": 0.0019,
326
+ "step": 4400
327
+ },
328
+ {
329
+ "epoch": 1.5126050420168067,
330
+ "grad_norm": 0.023098396137356758,
331
+ "learning_rate": 3.487731092436975e-05,
332
+ "loss": 0.0013,
333
+ "step": 4500
334
+ },
335
+ {
336
+ "epoch": 1.5462184873949578,
337
+ "grad_norm": 0.16987691819667816,
338
+ "learning_rate": 3.454117647058824e-05,
339
+ "loss": 0.0021,
340
+ "step": 4600
341
+ },
342
+ {
343
+ "epoch": 1.5798319327731094,
344
+ "grad_norm": 0.11584833264350891,
345
+ "learning_rate": 3.420504201680673e-05,
346
+ "loss": 0.0017,
347
+ "step": 4700
348
+ },
349
+ {
350
+ "epoch": 1.6134453781512605,
351
+ "grad_norm": 0.07376863062381744,
352
+ "learning_rate": 3.386890756302521e-05,
353
+ "loss": 0.0019,
354
+ "step": 4800
355
+ },
356
+ {
357
+ "epoch": 1.6470588235294117,
358
+ "grad_norm": 0.003893248736858368,
359
+ "learning_rate": 3.35327731092437e-05,
360
+ "loss": 0.0023,
361
+ "step": 4900
362
+ },
363
+ {
364
+ "epoch": 1.680672268907563,
365
+ "grad_norm": 0.11629138141870499,
366
+ "learning_rate": 3.3196638655462186e-05,
367
+ "loss": 0.002,
368
+ "step": 5000
369
+ },
370
+ {
371
+ "epoch": 1.7142857142857144,
372
+ "grad_norm": 0.09188009798526764,
373
+ "learning_rate": 3.286050420168068e-05,
374
+ "loss": 0.0017,
375
+ "step": 5100
376
+ },
377
+ {
378
+ "epoch": 1.7478991596638656,
379
+ "grad_norm": 0.11714625358581543,
380
+ "learning_rate": 3.2524369747899164e-05,
381
+ "loss": 0.0019,
382
+ "step": 5200
383
+ },
384
+ {
385
+ "epoch": 1.7815126050420167,
386
+ "grad_norm": 0.011511634103953838,
387
+ "learning_rate": 3.218823529411765e-05,
388
+ "loss": 0.0014,
389
+ "step": 5300
390
+ },
391
+ {
392
+ "epoch": 1.815126050420168,
393
+ "grad_norm": 0.025230418890714645,
394
+ "learning_rate": 3.1852100840336135e-05,
395
+ "loss": 0.0013,
396
+ "step": 5400
397
+ },
398
+ {
399
+ "epoch": 1.8487394957983194,
400
+ "grad_norm": 0.08280649781227112,
401
+ "learning_rate": 3.151596638655463e-05,
402
+ "loss": 0.0014,
403
+ "step": 5500
404
+ },
405
+ {
406
+ "epoch": 1.8823529411764706,
407
+ "grad_norm": 0.044068414717912674,
408
+ "learning_rate": 3.117983193277311e-05,
409
+ "loss": 0.0017,
410
+ "step": 5600
411
+ },
412
+ {
413
+ "epoch": 1.9159663865546217,
414
+ "grad_norm": 0.13378596305847168,
415
+ "learning_rate": 3.08436974789916e-05,
416
+ "loss": 0.0017,
417
+ "step": 5700
418
+ },
419
+ {
420
+ "epoch": 1.949579831932773,
421
+ "grad_norm": 0.08546321094036102,
422
+ "learning_rate": 3.0507563025210084e-05,
423
+ "loss": 0.0016,
424
+ "step": 5800
425
+ },
426
+ {
427
+ "epoch": 1.9831932773109244,
428
+ "grad_norm": 0.21406061947345734,
429
+ "learning_rate": 3.0171428571428572e-05,
430
+ "loss": 0.0017,
431
+ "step": 5900
432
+ },
433
+ {
434
+ "epoch": 2.0,
435
+ "eval_loss": 0.0019952901639044285,
436
+ "eval_runtime": 62.6295,
437
+ "eval_samples_per_second": 124.829,
438
+ "eval_steps_per_second": 7.808,
439
+ "step": 5950
440
+ },
441
+ {
442
+ "epoch": 2.0168067226890756,
443
+ "grad_norm": 0.12653397023677826,
444
+ "learning_rate": 2.983529411764706e-05,
445
+ "loss": 0.0012,
446
+ "step": 6000
447
+ },
448
+ {
449
+ "epoch": 2.0504201680672267,
450
+ "grad_norm": 0.021875187754631042,
451
+ "learning_rate": 2.949915966386555e-05,
452
+ "loss": 0.0011,
453
+ "step": 6100
454
+ },
455
+ {
456
+ "epoch": 2.0840336134453783,
457
+ "grad_norm": 0.02089652232825756,
458
+ "learning_rate": 2.9163025210084032e-05,
459
+ "loss": 0.0007,
460
+ "step": 6200
461
+ },
462
+ {
463
+ "epoch": 2.1176470588235294,
464
+ "grad_norm": 0.10652671009302139,
465
+ "learning_rate": 2.882689075630252e-05,
466
+ "loss": 0.0009,
467
+ "step": 6300
468
+ },
469
+ {
470
+ "epoch": 2.1512605042016806,
471
+ "grad_norm": 0.052155669778585434,
472
+ "learning_rate": 2.849075630252101e-05,
473
+ "loss": 0.001,
474
+ "step": 6400
475
+ },
476
+ {
477
+ "epoch": 2.184873949579832,
478
+ "grad_norm": 0.03689173236489296,
479
+ "learning_rate": 2.81546218487395e-05,
480
+ "loss": 0.0011,
481
+ "step": 6500
482
+ },
483
+ {
484
+ "epoch": 2.2184873949579833,
485
+ "grad_norm": 0.1453682780265808,
486
+ "learning_rate": 2.7818487394957988e-05,
487
+ "loss": 0.001,
488
+ "step": 6600
489
+ },
490
+ {
491
+ "epoch": 2.2521008403361344,
492
+ "grad_norm": 0.10738670825958252,
493
+ "learning_rate": 2.748235294117647e-05,
494
+ "loss": 0.0012,
495
+ "step": 6700
496
+ },
497
+ {
498
+ "epoch": 2.2857142857142856,
499
+ "grad_norm": 0.07871459424495697,
500
+ "learning_rate": 2.714621848739496e-05,
501
+ "loss": 0.0015,
502
+ "step": 6800
503
+ },
504
+ {
505
+ "epoch": 2.3193277310924367,
506
+ "grad_norm": 0.0664721354842186,
507
+ "learning_rate": 2.6810084033613448e-05,
508
+ "loss": 0.0011,
509
+ "step": 6900
510
+ },
511
+ {
512
+ "epoch": 2.3529411764705883,
513
+ "grad_norm": 0.0028383585158735514,
514
+ "learning_rate": 2.6473949579831937e-05,
515
+ "loss": 0.0012,
516
+ "step": 7000
517
+ },
518
+ {
519
+ "epoch": 2.3865546218487395,
520
+ "grad_norm": 0.06004633009433746,
521
+ "learning_rate": 2.613781512605042e-05,
522
+ "loss": 0.0011,
523
+ "step": 7100
524
+ },
525
+ {
526
+ "epoch": 2.4201680672268906,
527
+ "grad_norm": 0.1983553171157837,
528
+ "learning_rate": 2.5801680672268908e-05,
529
+ "loss": 0.0012,
530
+ "step": 7200
531
+ },
532
+ {
533
+ "epoch": 2.453781512605042,
534
+ "grad_norm": 0.10064397752285004,
535
+ "learning_rate": 2.5465546218487396e-05,
536
+ "loss": 0.0015,
537
+ "step": 7300
538
+ },
539
+ {
540
+ "epoch": 2.4873949579831933,
541
+ "grad_norm": 0.010177959688007832,
542
+ "learning_rate": 2.5129411764705885e-05,
543
+ "loss": 0.0012,
544
+ "step": 7400
545
+ },
546
+ {
547
+ "epoch": 2.5210084033613445,
548
+ "grad_norm": 0.0731307789683342,
549
+ "learning_rate": 2.479327731092437e-05,
550
+ "loss": 0.0013,
551
+ "step": 7500
552
+ },
553
+ {
554
+ "epoch": 2.5546218487394956,
555
+ "grad_norm": 0.015169462189078331,
556
+ "learning_rate": 2.445714285714286e-05,
557
+ "loss": 0.0014,
558
+ "step": 7600
559
+ },
560
+ {
561
+ "epoch": 2.588235294117647,
562
+ "grad_norm": 0.10259702801704407,
563
+ "learning_rate": 2.4121008403361345e-05,
564
+ "loss": 0.0009,
565
+ "step": 7700
566
+ },
567
+ {
568
+ "epoch": 2.6218487394957983,
569
+ "grad_norm": 0.11249709874391556,
570
+ "learning_rate": 2.3784873949579834e-05,
571
+ "loss": 0.0014,
572
+ "step": 7800
573
+ },
574
+ {
575
+ "epoch": 2.6554621848739495,
576
+ "grad_norm": 0.033951204270124435,
577
+ "learning_rate": 2.344873949579832e-05,
578
+ "loss": 0.0011,
579
+ "step": 7900
580
+ },
581
+ {
582
+ "epoch": 2.689075630252101,
583
+ "grad_norm": 0.07527955621480942,
584
+ "learning_rate": 2.311260504201681e-05,
585
+ "loss": 0.0011,
586
+ "step": 8000
587
+ },
588
+ {
589
+ "epoch": 2.722689075630252,
590
+ "grad_norm": 0.057067230343818665,
591
+ "learning_rate": 2.2776470588235297e-05,
592
+ "loss": 0.0008,
593
+ "step": 8100
594
+ },
595
+ {
596
+ "epoch": 2.7563025210084033,
597
+ "grad_norm": 0.028707563877105713,
598
+ "learning_rate": 2.2440336134453783e-05,
599
+ "loss": 0.0009,
600
+ "step": 8200
601
+ },
602
+ {
603
+ "epoch": 2.7899159663865545,
604
+ "grad_norm": 0.04694965109229088,
605
+ "learning_rate": 2.210420168067227e-05,
606
+ "loss": 0.0011,
607
+ "step": 8300
608
+ },
609
+ {
610
+ "epoch": 2.8235294117647056,
611
+ "grad_norm": 0.05145597457885742,
612
+ "learning_rate": 2.1768067226890757e-05,
613
+ "loss": 0.0009,
614
+ "step": 8400
615
+ },
616
+ {
617
+ "epoch": 2.857142857142857,
618
+ "grad_norm": 0.009829564020037651,
619
+ "learning_rate": 2.1431932773109246e-05,
620
+ "loss": 0.0007,
621
+ "step": 8500
622
+ },
623
+ {
624
+ "epoch": 2.8907563025210083,
625
+ "grad_norm": 0.017278891056776047,
626
+ "learning_rate": 2.109579831932773e-05,
627
+ "loss": 0.0007,
628
+ "step": 8600
629
+ },
630
+ {
631
+ "epoch": 2.92436974789916,
632
+ "grad_norm": 0.01149184163659811,
633
+ "learning_rate": 2.075966386554622e-05,
634
+ "loss": 0.0009,
635
+ "step": 8700
636
+ },
637
+ {
638
+ "epoch": 2.957983193277311,
639
+ "grad_norm": 0.024918945506215096,
640
+ "learning_rate": 2.042352941176471e-05,
641
+ "loss": 0.0007,
642
+ "step": 8800
643
+ },
644
+ {
645
+ "epoch": 2.991596638655462,
646
+ "grad_norm": 0.13216158747673035,
647
+ "learning_rate": 2.0087394957983195e-05,
648
+ "loss": 0.0007,
649
+ "step": 8900
650
+ },
651
+ {
652
+ "epoch": 3.0,
653
+ "eval_loss": 0.0018142093904316425,
654
+ "eval_runtime": 62.4784,
655
+ "eval_samples_per_second": 125.131,
656
+ "eval_steps_per_second": 7.827,
657
+ "step": 8925
658
+ },
659
+ {
660
+ "epoch": 3.0252100840336134,
661
+ "grad_norm": 0.006363982334733009,
662
+ "learning_rate": 1.9751260504201684e-05,
663
+ "loss": 0.0006,
664
+ "step": 9000
665
+ },
666
+ {
667
+ "epoch": 3.0588235294117645,
668
+ "grad_norm": 0.12958742678165436,
669
+ "learning_rate": 1.941512605042017e-05,
670
+ "loss": 0.0006,
671
+ "step": 9100
672
+ },
673
+ {
674
+ "epoch": 3.092436974789916,
675
+ "grad_norm": 0.00047478260239586234,
676
+ "learning_rate": 1.9078991596638658e-05,
677
+ "loss": 0.0004,
678
+ "step": 9200
679
+ },
680
+ {
681
+ "epoch": 3.1260504201680672,
682
+ "grad_norm": 0.1666508913040161,
683
+ "learning_rate": 1.8742857142857143e-05,
684
+ "loss": 0.0005,
685
+ "step": 9300
686
+ },
687
+ {
688
+ "epoch": 3.1596638655462184,
689
+ "grad_norm": 0.003551214002072811,
690
+ "learning_rate": 1.8406722689075632e-05,
691
+ "loss": 0.0007,
692
+ "step": 9400
693
+ },
694
+ {
695
+ "epoch": 3.19327731092437,
696
+ "grad_norm": 0.00195235013961792,
697
+ "learning_rate": 1.807058823529412e-05,
698
+ "loss": 0.0004,
699
+ "step": 9500
700
+ },
701
+ {
702
+ "epoch": 3.226890756302521,
703
+ "grad_norm": 0.004866756033152342,
704
+ "learning_rate": 1.7734453781512607e-05,
705
+ "loss": 0.0004,
706
+ "step": 9600
707
+ },
708
+ {
709
+ "epoch": 3.2605042016806722,
710
+ "grad_norm": 0.1616717129945755,
711
+ "learning_rate": 1.7398319327731096e-05,
712
+ "loss": 0.0005,
713
+ "step": 9700
714
+ },
715
+ {
716
+ "epoch": 3.2941176470588234,
717
+ "grad_norm": 0.1331503987312317,
718
+ "learning_rate": 1.706218487394958e-05,
719
+ "loss": 0.0005,
720
+ "step": 9800
721
+ },
722
+ {
723
+ "epoch": 3.327731092436975,
724
+ "grad_norm": 0.0034314363729208708,
725
+ "learning_rate": 1.672605042016807e-05,
726
+ "loss": 0.0006,
727
+ "step": 9900
728
+ },
729
+ {
730
+ "epoch": 3.361344537815126,
731
+ "grad_norm": 0.00794992409646511,
732
+ "learning_rate": 1.6389915966386555e-05,
733
+ "loss": 0.0003,
734
+ "step": 10000
735
+ },
736
+ {
737
+ "epoch": 3.3949579831932772,
738
+ "grad_norm": 0.003377394750714302,
739
+ "learning_rate": 1.6053781512605044e-05,
740
+ "loss": 0.0003,
741
+ "step": 10100
742
+ },
743
+ {
744
+ "epoch": 3.4285714285714284,
745
+ "grad_norm": 0.005109515506774187,
746
+ "learning_rate": 1.571764705882353e-05,
747
+ "loss": 0.0007,
748
+ "step": 10200
749
+ },
750
+ {
751
+ "epoch": 3.46218487394958,
752
+ "grad_norm": 0.08545704185962677,
753
+ "learning_rate": 1.538151260504202e-05,
754
+ "loss": 0.0007,
755
+ "step": 10300
756
+ },
757
+ {
758
+ "epoch": 3.495798319327731,
759
+ "grad_norm": 0.0017083921702578664,
760
+ "learning_rate": 1.5045378151260506e-05,
761
+ "loss": 0.0005,
762
+ "step": 10400
763
+ },
764
+ {
765
+ "epoch": 3.5294117647058822,
766
+ "grad_norm": 0.011194012127816677,
767
+ "learning_rate": 1.4709243697478991e-05,
768
+ "loss": 0.0007,
769
+ "step": 10500
770
+ },
771
+ {
772
+ "epoch": 3.5630252100840334,
773
+ "grad_norm": 0.20523475110530853,
774
+ "learning_rate": 1.437310924369748e-05,
775
+ "loss": 0.0003,
776
+ "step": 10600
777
+ },
778
+ {
779
+ "epoch": 3.596638655462185,
780
+ "grad_norm": 0.06024377420544624,
781
+ "learning_rate": 1.4036974789915966e-05,
782
+ "loss": 0.0007,
783
+ "step": 10700
784
+ },
785
+ {
786
+ "epoch": 3.630252100840336,
787
+ "grad_norm": 0.11052771657705307,
788
+ "learning_rate": 1.3700840336134455e-05,
789
+ "loss": 0.0005,
790
+ "step": 10800
791
+ },
792
+ {
793
+ "epoch": 3.6638655462184873,
794
+ "grad_norm": 0.2950091063976288,
795
+ "learning_rate": 1.3364705882352943e-05,
796
+ "loss": 0.0005,
797
+ "step": 10900
798
+ },
799
+ {
800
+ "epoch": 3.697478991596639,
801
+ "grad_norm": 0.016148166730999947,
802
+ "learning_rate": 1.3028571428571429e-05,
803
+ "loss": 0.0005,
804
+ "step": 11000
805
+ },
806
+ {
807
+ "epoch": 3.73109243697479,
808
+ "grad_norm": 0.15638086199760437,
809
+ "learning_rate": 1.2692436974789918e-05,
810
+ "loss": 0.0005,
811
+ "step": 11100
812
+ },
813
+ {
814
+ "epoch": 3.764705882352941,
815
+ "grad_norm": 0.00034360421705059707,
816
+ "learning_rate": 1.2356302521008405e-05,
817
+ "loss": 0.0005,
818
+ "step": 11200
819
+ },
820
+ {
821
+ "epoch": 3.7983193277310923,
822
+ "grad_norm": 0.026777559891343117,
823
+ "learning_rate": 1.2020168067226892e-05,
824
+ "loss": 0.0004,
825
+ "step": 11300
826
+ },
827
+ {
828
+ "epoch": 3.831932773109244,
829
+ "grad_norm": 0.038060788065195084,
830
+ "learning_rate": 1.168403361344538e-05,
831
+ "loss": 0.0005,
832
+ "step": 11400
833
+ },
834
+ {
835
+ "epoch": 3.865546218487395,
836
+ "grad_norm": 0.003074630396440625,
837
+ "learning_rate": 1.1347899159663867e-05,
838
+ "loss": 0.0004,
839
+ "step": 11500
840
+ },
841
+ {
842
+ "epoch": 3.899159663865546,
843
+ "grad_norm": 0.0007972168968990445,
844
+ "learning_rate": 1.1011764705882354e-05,
845
+ "loss": 0.0004,
846
+ "step": 11600
847
+ },
848
+ {
849
+ "epoch": 3.9327731092436977,
850
+ "grad_norm": 0.011879481375217438,
851
+ "learning_rate": 1.0675630252100841e-05,
852
+ "loss": 0.0005,
853
+ "step": 11700
854
+ },
855
+ {
856
+ "epoch": 3.966386554621849,
857
+ "grad_norm": 0.0010381491156294942,
858
+ "learning_rate": 1.0339495798319328e-05,
859
+ "loss": 0.0004,
860
+ "step": 11800
861
+ },
862
+ {
863
+ "epoch": 4.0,
864
+ "grad_norm": 0.0010164816631004214,
865
+ "learning_rate": 1.0003361344537815e-05,
866
+ "loss": 0.0003,
867
+ "step": 11900
868
+ },
869
+ {
870
+ "epoch": 4.0,
871
+ "eval_loss": 0.0016061929054558277,
872
+ "eval_runtime": 62.4482,
873
+ "eval_samples_per_second": 125.192,
874
+ "eval_steps_per_second": 7.83,
875
+ "step": 11900
876
+ }
877
+ ],
878
+ "logging_steps": 100,
879
+ "max_steps": 14875,
880
+ "num_input_tokens_seen": 0,
881
+ "num_train_epochs": 5,
882
+ "save_steps": 500,
883
+ "stateful_callbacks": {
884
+ "TrainerControl": {
885
+ "args": {
886
+ "should_epoch_stop": false,
887
+ "should_evaluate": false,
888
+ "should_log": false,
889
+ "should_save": true,
890
+ "should_training_stop": false
891
+ },
892
+ "attributes": {}
893
+ }
894
+ },
895
+ "total_flos": 1.71886155388416e+16,
896
+ "train_batch_size": 16,
897
+ "trial_name": null,
898
+ "trial_params": null
899
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9664e2d990115a6fbe2af32d09661958f7cb1ee1b261c2a1164dff293dcfdf04
3
+ size 5368