AlekseyKorshuk commited on
Commit
a625fe5
1 Parent(s): 7c21bbe

Training in progress, step 82

Browse files
Files changed (4) hide show
  1. config.json +20 -17
  2. tokenizer.json +0 -0
  3. tokenizer_config.json +78 -126
  4. training_args.bin +1 -1
config.json CHANGED
@@ -1,31 +1,34 @@
1
  {
2
- "_name_or_path": "stabilityai/stable-code-3b",
3
  "architectures": [
4
- "StableLMEpochForCausalLM"
5
  ],
 
6
  "auto_map": {
7
- "AutoConfig": "configuration_stablelm_epoch.StableLMEpochConfig",
8
- "AutoModelForCausalLM": "modeling_stablelm_epoch.StableLMEpochForCausalLM"
9
  },
10
- "bos_token_id": 0,
11
- "eos_token_id": 0,
12
- "hidden_act": "silu",
 
13
  "hidden_size": 2560,
14
  "initializer_range": 0.02,
15
- "intermediate_size": 6912,
16
- "max_position_embeddings": 16384,
17
- "model_type": "stablelm_epoch",
18
- "norm_eps": 1e-05,
19
  "num_attention_heads": 32,
20
- "num_heads": 32,
21
  "num_hidden_layers": 32,
22
  "num_key_value_heads": 32,
23
- "rope_pct": 0.25,
24
- "rope_theta": 1000000,
25
- "rotary_scaling_factor": 1.0,
 
 
26
  "tie_word_embeddings": false,
27
- "torch_dtype": "bfloat16",
28
  "transformers_version": "4.37.0.dev0",
29
  "use_cache": false,
30
- "vocab_size": 50304
31
  }
 
1
  {
2
+ "_name_or_path": "microsoft/phi-2",
3
  "architectures": [
4
+ "PhiForCausalLM"
5
  ],
6
+ "attention_dropout": 0.0,
7
  "auto_map": {
8
+ "AutoConfig": "microsoft/phi-2--configuration_phi.PhiConfig",
9
+ "AutoModelForCausalLM": "microsoft/phi-2--modeling_phi.PhiForCausalLM"
10
  },
11
+ "bos_token_id": null,
12
+ "embd_pdrop": 0.0,
13
+ "eos_token_id": null,
14
+ "hidden_act": "gelu_new",
15
  "hidden_size": 2560,
16
  "initializer_range": 0.02,
17
+ "intermediate_size": 10240,
18
+ "layer_norm_eps": 1e-05,
19
+ "max_position_embeddings": 2048,
20
+ "model_type": "phi",
21
  "num_attention_heads": 32,
 
22
  "num_hidden_layers": 32,
23
  "num_key_value_heads": 32,
24
+ "partial_rotary_factor": 0.4,
25
+ "qk_layernorm": false,
26
+ "resid_pdrop": 0.1,
27
+ "rope_scaling": null,
28
+ "rope_theta": 10000.0,
29
  "tie_word_embeddings": false,
30
+ "torch_dtype": "float16",
31
  "transformers_version": "4.37.0.dev0",
32
  "use_cache": false,
33
+ "vocab_size": 51200
34
  }
tokenizer.json CHANGED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json CHANGED
@@ -1,7 +1,7 @@
1
  {
2
  "add_prefix_space": false,
3
  "added_tokens_decoder": {
4
- "0": {
5
  "content": "<|endoftext|>",
6
  "lstrip": false,
7
  "normalized": false,
@@ -9,40 +9,8 @@
9
  "single_word": false,
10
  "special": true
11
  },
12
- "1": {
13
- "content": "<|padding|>",
14
- "lstrip": false,
15
- "normalized": false,
16
- "rstrip": false,
17
- "single_word": false,
18
- "special": true
19
- },
20
- "50254": {
21
- "content": " ",
22
- "lstrip": false,
23
- "normalized": true,
24
- "rstrip": false,
25
- "single_word": false,
26
- "special": false
27
- },
28
- "50255": {
29
- "content": " ",
30
- "lstrip": false,
31
- "normalized": true,
32
- "rstrip": false,
33
- "single_word": false,
34
- "special": false
35
- },
36
- "50256": {
37
- "content": " ",
38
- "lstrip": false,
39
- "normalized": true,
40
- "rstrip": false,
41
- "single_word": false,
42
- "special": false
43
- },
44
  "50257": {
45
- "content": " ",
46
  "lstrip": false,
47
  "normalized": true,
48
  "rstrip": false,
@@ -50,7 +18,7 @@
50
  "special": false
51
  },
52
  "50258": {
53
- "content": " ",
54
  "lstrip": false,
55
  "normalized": true,
56
  "rstrip": false,
@@ -58,7 +26,7 @@
58
  "special": false
59
  },
60
  "50259": {
61
- "content": " ",
62
  "lstrip": false,
63
  "normalized": true,
64
  "rstrip": false,
@@ -66,7 +34,7 @@
66
  "special": false
67
  },
68
  "50260": {
69
- "content": " ",
70
  "lstrip": false,
71
  "normalized": true,
72
  "rstrip": false,
@@ -74,7 +42,7 @@
74
  "special": false
75
  },
76
  "50261": {
77
- "content": " ",
78
  "lstrip": false,
79
  "normalized": true,
80
  "rstrip": false,
@@ -82,7 +50,7 @@
82
  "special": false
83
  },
84
  "50262": {
85
- "content": " ",
86
  "lstrip": false,
87
  "normalized": true,
88
  "rstrip": false,
@@ -90,7 +58,7 @@
90
  "special": false
91
  },
92
  "50263": {
93
- "content": " ",
94
  "lstrip": false,
95
  "normalized": true,
96
  "rstrip": false,
@@ -98,7 +66,7 @@
98
  "special": false
99
  },
100
  "50264": {
101
- "content": " ",
102
  "lstrip": false,
103
  "normalized": true,
104
  "rstrip": false,
@@ -106,7 +74,7 @@
106
  "special": false
107
  },
108
  "50265": {
109
- "content": " ",
110
  "lstrip": false,
111
  "normalized": true,
112
  "rstrip": false,
@@ -114,7 +82,7 @@
114
  "special": false
115
  },
116
  "50266": {
117
- "content": " ",
118
  "lstrip": false,
119
  "normalized": true,
120
  "rstrip": false,
@@ -122,7 +90,7 @@
122
  "special": false
123
  },
124
  "50267": {
125
- "content": " ",
126
  "lstrip": false,
127
  "normalized": true,
128
  "rstrip": false,
@@ -130,7 +98,7 @@
130
  "special": false
131
  },
132
  "50268": {
133
- "content": " ",
134
  "lstrip": false,
135
  "normalized": true,
136
  "rstrip": false,
@@ -138,7 +106,7 @@
138
  "special": false
139
  },
140
  "50269": {
141
- "content": " ",
142
  "lstrip": false,
143
  "normalized": true,
144
  "rstrip": false,
@@ -146,7 +114,7 @@
146
  "special": false
147
  },
148
  "50270": {
149
- "content": " ",
150
  "lstrip": false,
151
  "normalized": true,
152
  "rstrip": false,
@@ -154,7 +122,7 @@
154
  "special": false
155
  },
156
  "50271": {
157
- "content": " ",
158
  "lstrip": false,
159
  "normalized": true,
160
  "rstrip": false,
@@ -162,7 +130,7 @@
162
  "special": false
163
  },
164
  "50272": {
165
- "content": " ",
166
  "lstrip": false,
167
  "normalized": true,
168
  "rstrip": false,
@@ -170,7 +138,7 @@
170
  "special": false
171
  },
172
  "50273": {
173
- "content": " ",
174
  "lstrip": false,
175
  "normalized": true,
176
  "rstrip": false,
@@ -178,7 +146,7 @@
178
  "special": false
179
  },
180
  "50274": {
181
- "content": " ",
182
  "lstrip": false,
183
  "normalized": true,
184
  "rstrip": false,
@@ -186,7 +154,7 @@
186
  "special": false
187
  },
188
  "50275": {
189
- "content": " ",
190
  "lstrip": false,
191
  "normalized": true,
192
  "rstrip": false,
@@ -194,7 +162,7 @@
194
  "special": false
195
  },
196
  "50276": {
197
- "content": " ",
198
  "lstrip": false,
199
  "normalized": true,
200
  "rstrip": false,
@@ -202,166 +170,150 @@
202
  "special": false
203
  },
204
  "50277": {
205
- "content": "<fim_prefix>",
206
  "lstrip": false,
207
- "normalized": false,
208
  "rstrip": false,
209
  "single_word": false,
210
- "special": true
211
  },
212
  "50278": {
213
- "content": "<fim_middle>",
214
  "lstrip": false,
215
- "normalized": false,
216
  "rstrip": false,
217
  "single_word": false,
218
- "special": true
219
  },
220
  "50279": {
221
- "content": "<fim_suffix>",
222
  "lstrip": false,
223
- "normalized": false,
224
  "rstrip": false,
225
  "single_word": false,
226
- "special": true
227
  },
228
  "50280": {
229
- "content": "<fim_pad>",
230
  "lstrip": false,
231
- "normalized": false,
232
  "rstrip": false,
233
  "single_word": false,
234
- "special": true
235
  },
236
  "50281": {
237
- "content": "<filename>",
238
  "lstrip": false,
239
- "normalized": false,
240
  "rstrip": false,
241
  "single_word": false,
242
- "special": true
243
  },
244
  "50282": {
245
- "content": "<gh_stars>",
246
  "lstrip": false,
247
- "normalized": false,
248
  "rstrip": false,
249
  "single_word": false,
250
- "special": true
251
  },
252
  "50283": {
253
- "content": "<issue_start>",
254
  "lstrip": false,
255
- "normalized": false,
256
  "rstrip": false,
257
  "single_word": false,
258
- "special": true
259
  },
260
  "50284": {
261
- "content": "<issue_comment>",
262
  "lstrip": false,
263
- "normalized": false,
264
  "rstrip": false,
265
  "single_word": false,
266
- "special": true
267
  },
268
  "50285": {
269
- "content": "<issue_closed>",
270
  "lstrip": false,
271
- "normalized": false,
272
  "rstrip": false,
273
  "single_word": false,
274
- "special": true
275
  },
276
  "50286": {
277
- "content": "<jupyter_start>",
278
  "lstrip": false,
279
- "normalized": false,
280
  "rstrip": false,
281
  "single_word": false,
282
- "special": true
283
  },
284
  "50287": {
285
- "content": "<jupyter_text>",
286
  "lstrip": false,
287
- "normalized": false,
288
  "rstrip": false,
289
  "single_word": false,
290
- "special": true
291
  },
292
  "50288": {
293
- "content": "<jupyter_code>",
294
  "lstrip": false,
295
- "normalized": false,
296
  "rstrip": false,
297
  "single_word": false,
298
- "special": true
299
  },
300
  "50289": {
301
- "content": "<jupyter_output>",
302
  "lstrip": false,
303
- "normalized": false,
304
  "rstrip": false,
305
  "single_word": false,
306
- "special": true
307
  },
308
  "50290": {
309
- "content": "<empty_output>",
310
  "lstrip": false,
311
- "normalized": false,
312
  "rstrip": false,
313
  "single_word": false,
314
- "special": true
315
  },
316
  "50291": {
317
- "content": "<commit_before>",
318
  "lstrip": false,
319
- "normalized": false,
320
  "rstrip": false,
321
  "single_word": false,
322
- "special": true
323
  },
324
  "50292": {
325
- "content": "<commit_msg>",
326
  "lstrip": false,
327
- "normalized": false,
328
  "rstrip": false,
329
  "single_word": false,
330
- "special": true
331
  },
332
  "50293": {
333
- "content": "<commit_after>",
334
  "lstrip": false,
335
- "normalized": false,
336
  "rstrip": false,
337
  "single_word": false,
338
- "special": true
339
  },
340
  "50294": {
341
- "content": "<reponame>",
342
  "lstrip": false,
343
- "normalized": false,
344
  "rstrip": false,
345
  "single_word": false,
346
- "special": true
347
  },
348
  "50295": {
349
- "content": "<repo_continuation>",
350
- "lstrip": false,
351
- "normalized": false,
352
- "rstrip": false,
353
- "single_word": false,
354
- "special": true
355
- },
356
- "50296": {
357
- "content": "[PAD]",
358
- "lstrip": false,
359
- "normalized": false,
360
- "rstrip": false,
361
- "single_word": false,
362
- "special": true
363
- },
364
- "50297": {
365
  "content": "<|im_end|>",
366
  "lstrip": false,
367
  "normalized": false,
@@ -369,7 +321,7 @@
369
  "single_word": false,
370
  "special": true
371
  },
372
- "50298": {
373
  "content": "<|im_start|>",
374
  "lstrip": false,
375
  "normalized": false,
@@ -381,8 +333,8 @@
381
  "bos_token": "<|endoftext|>",
382
  "clean_up_tokenization_spaces": true,
383
  "eos_token": "<|im_end|>",
384
- "model_max_length": 1000000000000000019884624838656,
385
  "pad_token": "<|endoftext|>",
386
- "tokenizer_class": "GPTNeoXTokenizer",
387
  "unk_token": "<|endoftext|>"
388
  }
 
1
  {
2
  "add_prefix_space": false,
3
  "added_tokens_decoder": {
4
+ "50256": {
5
  "content": "<|endoftext|>",
6
  "lstrip": false,
7
  "normalized": false,
 
9
  "single_word": false,
10
  "special": true
11
  },
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
12
  "50257": {
13
+ "content": " ",
14
  "lstrip": false,
15
  "normalized": true,
16
  "rstrip": false,
 
18
  "special": false
19
  },
20
  "50258": {
21
+ "content": " ",
22
  "lstrip": false,
23
  "normalized": true,
24
  "rstrip": false,
 
26
  "special": false
27
  },
28
  "50259": {
29
+ "content": " ",
30
  "lstrip": false,
31
  "normalized": true,
32
  "rstrip": false,
 
34
  "special": false
35
  },
36
  "50260": {
37
+ "content": " ",
38
  "lstrip": false,
39
  "normalized": true,
40
  "rstrip": false,
 
42
  "special": false
43
  },
44
  "50261": {
45
+ "content": " ",
46
  "lstrip": false,
47
  "normalized": true,
48
  "rstrip": false,
 
50
  "special": false
51
  },
52
  "50262": {
53
+ "content": " ",
54
  "lstrip": false,
55
  "normalized": true,
56
  "rstrip": false,
 
58
  "special": false
59
  },
60
  "50263": {
61
+ "content": " ",
62
  "lstrip": false,
63
  "normalized": true,
64
  "rstrip": false,
 
66
  "special": false
67
  },
68
  "50264": {
69
+ "content": " ",
70
  "lstrip": false,
71
  "normalized": true,
72
  "rstrip": false,
 
74
  "special": false
75
  },
76
  "50265": {
77
+ "content": " ",
78
  "lstrip": false,
79
  "normalized": true,
80
  "rstrip": false,
 
82
  "special": false
83
  },
84
  "50266": {
85
+ "content": " ",
86
  "lstrip": false,
87
  "normalized": true,
88
  "rstrip": false,
 
90
  "special": false
91
  },
92
  "50267": {
93
+ "content": " ",
94
  "lstrip": false,
95
  "normalized": true,
96
  "rstrip": false,
 
98
  "special": false
99
  },
100
  "50268": {
101
+ "content": " ",
102
  "lstrip": false,
103
  "normalized": true,
104
  "rstrip": false,
 
106
  "special": false
107
  },
108
  "50269": {
109
+ "content": " ",
110
  "lstrip": false,
111
  "normalized": true,
112
  "rstrip": false,
 
114
  "special": false
115
  },
116
  "50270": {
117
+ "content": " ",
118
  "lstrip": false,
119
  "normalized": true,
120
  "rstrip": false,
 
122
  "special": false
123
  },
124
  "50271": {
125
+ "content": " ",
126
  "lstrip": false,
127
  "normalized": true,
128
  "rstrip": false,
 
130
  "special": false
131
  },
132
  "50272": {
133
+ "content": " ",
134
  "lstrip": false,
135
  "normalized": true,
136
  "rstrip": false,
 
138
  "special": false
139
  },
140
  "50273": {
141
+ "content": " ",
142
  "lstrip": false,
143
  "normalized": true,
144
  "rstrip": false,
 
146
  "special": false
147
  },
148
  "50274": {
149
+ "content": " ",
150
  "lstrip": false,
151
  "normalized": true,
152
  "rstrip": false,
 
154
  "special": false
155
  },
156
  "50275": {
157
+ "content": " ",
158
  "lstrip": false,
159
  "normalized": true,
160
  "rstrip": false,
 
162
  "special": false
163
  },
164
  "50276": {
165
+ "content": " ",
166
  "lstrip": false,
167
  "normalized": true,
168
  "rstrip": false,
 
170
  "special": false
171
  },
172
  "50277": {
173
+ "content": " ",
174
  "lstrip": false,
175
+ "normalized": true,
176
  "rstrip": false,
177
  "single_word": false,
178
+ "special": false
179
  },
180
  "50278": {
181
+ "content": " ",
182
  "lstrip": false,
183
+ "normalized": true,
184
  "rstrip": false,
185
  "single_word": false,
186
+ "special": false
187
  },
188
  "50279": {
189
+ "content": " ",
190
  "lstrip": false,
191
+ "normalized": true,
192
  "rstrip": false,
193
  "single_word": false,
194
+ "special": false
195
  },
196
  "50280": {
197
+ "content": " ",
198
  "lstrip": false,
199
+ "normalized": true,
200
  "rstrip": false,
201
  "single_word": false,
202
+ "special": false
203
  },
204
  "50281": {
205
+ "content": " ",
206
  "lstrip": false,
207
+ "normalized": true,
208
  "rstrip": false,
209
  "single_word": false,
210
+ "special": false
211
  },
212
  "50282": {
213
+ "content": " ",
214
  "lstrip": false,
215
+ "normalized": true,
216
  "rstrip": false,
217
  "single_word": false,
218
+ "special": false
219
  },
220
  "50283": {
221
+ "content": " ",
222
  "lstrip": false,
223
+ "normalized": true,
224
  "rstrip": false,
225
  "single_word": false,
226
+ "special": false
227
  },
228
  "50284": {
229
+ "content": " ",
230
  "lstrip": false,
231
+ "normalized": true,
232
  "rstrip": false,
233
  "single_word": false,
234
+ "special": false
235
  },
236
  "50285": {
237
+ "content": " ",
238
  "lstrip": false,
239
+ "normalized": true,
240
  "rstrip": false,
241
  "single_word": false,
242
+ "special": false
243
  },
244
  "50286": {
245
+ "content": " ",
246
  "lstrip": false,
247
+ "normalized": true,
248
  "rstrip": false,
249
  "single_word": false,
250
+ "special": false
251
  },
252
  "50287": {
253
+ "content": "\t\t\t\t\t\t\t\t\t",
254
  "lstrip": false,
255
+ "normalized": true,
256
  "rstrip": false,
257
  "single_word": false,
258
+ "special": false
259
  },
260
  "50288": {
261
+ "content": "\t\t\t\t\t\t\t\t",
262
  "lstrip": false,
263
+ "normalized": true,
264
  "rstrip": false,
265
  "single_word": false,
266
+ "special": false
267
  },
268
  "50289": {
269
+ "content": "\t\t\t\t\t\t\t",
270
  "lstrip": false,
271
+ "normalized": true,
272
  "rstrip": false,
273
  "single_word": false,
274
+ "special": false
275
  },
276
  "50290": {
277
+ "content": "\t\t\t\t\t\t",
278
  "lstrip": false,
279
+ "normalized": true,
280
  "rstrip": false,
281
  "single_word": false,
282
+ "special": false
283
  },
284
  "50291": {
285
+ "content": "\t\t\t\t\t",
286
  "lstrip": false,
287
+ "normalized": true,
288
  "rstrip": false,
289
  "single_word": false,
290
+ "special": false
291
  },
292
  "50292": {
293
+ "content": "\t\t\t\t",
294
  "lstrip": false,
295
+ "normalized": true,
296
  "rstrip": false,
297
  "single_word": false,
298
+ "special": false
299
  },
300
  "50293": {
301
+ "content": "\t\t\t",
302
  "lstrip": false,
303
+ "normalized": true,
304
  "rstrip": false,
305
  "single_word": false,
306
+ "special": false
307
  },
308
  "50294": {
309
+ "content": "\t\t",
310
  "lstrip": false,
311
+ "normalized": true,
312
  "rstrip": false,
313
  "single_word": false,
314
+ "special": false
315
  },
316
  "50295": {
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
317
  "content": "<|im_end|>",
318
  "lstrip": false,
319
  "normalized": false,
 
321
  "single_word": false,
322
  "special": true
323
  },
324
+ "50296": {
325
  "content": "<|im_start|>",
326
  "lstrip": false,
327
  "normalized": false,
 
333
  "bos_token": "<|endoftext|>",
334
  "clean_up_tokenization_spaces": true,
335
  "eos_token": "<|im_end|>",
336
+ "model_max_length": 2048,
337
  "pad_token": "<|endoftext|>",
338
+ "tokenizer_class": "CodeGenTokenizer",
339
  "unk_token": "<|endoftext|>"
340
  }
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:965952c63f3d85ba9b126050b91a41ef727cd00c59a89e0f6e549d31ead26108
3
  size 4859
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9d49bfb580919e1b3b8b2da718ed7ac52ea5d8f89515934236bec46c7e0d89b7
3
  size 4859