Mhair commited on
Commit
f833e6d
·
verified ·
1 Parent(s): a147862

Run 5. Outer Step 32. Inner Step 2.

Browse files
Files changed (4) hide show
  1. config.json +70 -82
  2. generation_config.json +7 -0
  3. inner_optimizer.pt +2 -2
  4. model.safetensors +2 -2
config.json CHANGED
@@ -1,13 +1,12 @@
1
  {
2
- "_name_or_path": "distributed/optimized-gpt2-1b",
3
- "activation_function": "gelu_new",
4
  "all_reduce_scores": {
5
  "0": "NON_PARTICIPATING",
6
  "1": "SUCCESS",
7
  "10": "NON_PARTICIPATING",
8
  "100": "NON_PARTICIPATING",
9
  "101": "NON_PARTICIPATING",
10
- "102": "SUCCESS",
11
  "103": "NON_PARTICIPATING",
12
  "104": "NON_PARTICIPATING",
13
  "105": "NON_PARTICIPATING",
@@ -26,26 +25,26 @@
26
  "117": "NON_PARTICIPATING",
27
  "118": "NON_PARTICIPATING",
28
  "119": "NON_PARTICIPATING",
29
- "12": "SUCCESS",
30
  "120": "NON_PARTICIPATING",
31
  "121": "NON_PARTICIPATING",
32
  "122": "NON_PARTICIPATING",
33
  "123": "NON_PARTICIPATING",
34
- "124": "SUCCESS",
35
  "125": "NON_PARTICIPATING",
36
  "126": "NON_PARTICIPATING",
37
  "127": "NON_PARTICIPATING",
38
  "128": "NON_PARTICIPATING",
39
  "129": "NON_PARTICIPATING",
40
  "13": "SUCCESS",
41
- "130": "SUCCESS",
42
  "131": "NON_PARTICIPATING",
43
  "132": "NON_PARTICIPATING",
44
  "133": "NON_PARTICIPATING",
45
- "134": "NON_PARTICIPATING",
46
  "135": "NON_PARTICIPATING",
47
  "136": "NON_PARTICIPATING",
48
- "137": "SUCCESS",
49
  "138": "NON_PARTICIPATING",
50
  "139": "NON_PARTICIPATING",
51
  "14": "SUCCESS",
@@ -56,14 +55,14 @@
56
  "144": "NON_PARTICIPATING",
57
  "145": "NON_PARTICIPATING",
58
  "146": "NON_PARTICIPATING",
59
- "147": "SUCCESS",
60
  "148": "NON_PARTICIPATING",
61
  "149": "NON_PARTICIPATING",
62
  "15": "SUCCESS",
63
- "150": "SUCCESS",
64
  "151": "NON_PARTICIPATING",
65
  "152": "NON_PARTICIPATING",
66
- "153": "SUCCESS",
67
  "154": "NON_PARTICIPATING",
68
  "155": "NON_PARTICIPATING",
69
  "156": "NON_PARTICIPATING",
@@ -73,7 +72,7 @@
73
  "16": "SUCCESS",
74
  "160": "NON_PARTICIPATING",
75
  "161": "NON_PARTICIPATING",
76
- "162": "SUCCESS",
77
  "163": "NON_PARTICIPATING",
78
  "164": "NON_PARTICIPATING",
79
  "165": "NON_PARTICIPATING",
@@ -92,14 +91,14 @@
92
  "177": "NON_PARTICIPATING",
93
  "178": "NON_PARTICIPATING",
94
  "179": "NON_PARTICIPATING",
95
- "18": "SUCCESS",
96
  "180": "NON_PARTICIPATING",
97
  "181": "NON_PARTICIPATING",
98
  "182": "NON_PARTICIPATING",
99
  "183": "NON_PARTICIPATING",
100
  "184": "NON_PARTICIPATING",
101
- "185": "SUCCESS",
102
- "186": "SUCCESS",
103
  "187": "NON_PARTICIPATING",
104
  "188": "NON_PARTICIPATING",
105
  "189": "NON_PARTICIPATING",
@@ -116,12 +115,12 @@
116
  "199": "NON_PARTICIPATING",
117
  "2": "SUCCESS",
118
  "20": "SUCCESS",
119
- "200": "NON_PARTICIPATING",
120
  "201": "NON_PARTICIPATING",
121
  "202": "NON_PARTICIPATING",
122
- "203": "SUCCESS",
123
- "204": "SUCCESS",
124
- "205": "SUCCESS",
125
  "206": "NON_PARTICIPATING",
126
  "207": "NON_PARTICIPATING",
127
  "208": "SUCCESS",
@@ -135,7 +134,7 @@
135
  "215": "NON_PARTICIPATING",
136
  "216": "NON_PARTICIPATING",
137
  "217": "NON_PARTICIPATING",
138
- "218": "NON_PARTICIPATING",
139
  "219": "NON_PARTICIPATING",
140
  "22": "SUCCESS",
141
  "220": "NON_PARTICIPATING",
@@ -147,10 +146,10 @@
147
  "226": "NON_PARTICIPATING",
148
  "227": "NON_PARTICIPATING",
149
  "228": "NON_PARTICIPATING",
150
- "229": "SUCCESS",
151
- "23": "SUCCESS",
152
  "230": "NON_PARTICIPATING",
153
- "231": "SUCCESS",
154
  "232": "NON_PARTICIPATING",
155
  "233": "NON_PARTICIPATING",
156
  "234": "NON_PARTICIPATING",
@@ -164,20 +163,20 @@
164
  "241": "NON_PARTICIPATING",
165
  "242": "NON_PARTICIPATING",
166
  "243": "NON_PARTICIPATING",
167
- "244": "SUCCESS",
168
  "245": "NON_PARTICIPATING",
169
- "246": "NON_PARTICIPATING",
170
  "247": "NON_PARTICIPATING",
171
  "248": "NON_PARTICIPATING",
172
- "249": "NON_PARTICIPATING",
173
  "25": "SUCCESS",
174
  "250": "NON_PARTICIPATING",
175
  "251": "NON_PARTICIPATING",
176
- "252": "SUCCESS",
177
  "253": "NON_PARTICIPATING",
178
  "254": "NON_PARTICIPATING",
179
  "255": "NON_PARTICIPATING",
180
- "26": "SUCCESS",
181
  "27": "SUCCESS",
182
  "28": "SUCCESS",
183
  "29": "SUCCESS",
@@ -186,7 +185,7 @@
186
  "31": "SUCCESS",
187
  "32": "SUCCESS",
188
  "33": "NON_PARTICIPATING",
189
- "34": "SUCCESS",
190
  "35": "SUCCESS",
191
  "36": "SUCCESS",
192
  "37": "SUCCESS",
@@ -199,12 +198,12 @@
199
  "43": "SUCCESS",
200
  "44": "SUCCESS",
201
  "45": "SUCCESS",
202
- "46": "NON_PARTICIPATING",
203
- "47": "NON_PARTICIPATING",
204
- "48": "NON_PARTICIPATING",
205
  "49": "SUCCESS",
206
  "5": "SUCCESS",
207
- "50": "SUCCESS",
208
  "51": "SUCCESS",
209
  "52": "SUCCESS",
210
  "53": "SUCCESS",
@@ -212,22 +211,22 @@
212
  "55": "SUCCESS",
213
  "56": "SUCCESS",
214
  "57": "SUCCESS",
215
- "58": "NON_PARTICIPATING",
216
  "59": "SUCCESS",
217
- "6": "NON_PARTICIPATING",
218
  "60": "SUCCESS",
219
- "61": "NON_PARTICIPATING",
220
  "62": "SUCCESS",
221
  "63": "SUCCESS",
222
- "64": "NON_PARTICIPATING",
223
  "65": "SUCCESS",
224
  "66": "SUCCESS",
225
  "67": "SUCCESS",
226
  "68": "SUCCESS",
227
  "69": "SUCCESS",
228
  "7": "SUCCESS",
229
- "70": "NON_PARTICIPATING",
230
- "71": "NON_PARTICIPATING",
231
  "72": "SUCCESS",
232
  "73": "NON_PARTICIPATING",
233
  "74": "SUCCESS",
@@ -235,9 +234,9 @@
235
  "76": "SUCCESS",
236
  "77": "SUCCESS",
237
  "78": "SUCCESS",
238
- "79": "NON_PARTICIPATING",
239
  "8": "SUCCESS",
240
- "80": "SUCCESS",
241
  "81": "SUCCESS",
242
  "82": "SUCCESS",
243
  "83": "NON_PARTICIPATING",
@@ -245,61 +244,50 @@
245
  "85": "SUCCESS",
246
  "86": "SUCCESS",
247
  "87": "SUCCESS",
248
- "88": "SUCCESS",
249
  "89": "SUCCESS",
250
- "9": "NON_PARTICIPATING",
251
- "90": "SUCCESS",
252
  "91": "SUCCESS",
253
  "92": "SUCCESS",
254
- "93": "NON_PARTICIPATING",
255
  "94": "SUCCESS",
256
  "95": "SUCCESS",
257
- "96": "SUCCESS",
258
  "97": "SUCCESS",
259
  "98": "SUCCESS",
260
  "99": "SUCCESS"
261
  },
262
  "architectures": [
263
- "GPTOptim"
264
  ],
265
- "attn_pdrop": 0.1,
266
- "auto_map": {
267
- "AutoConfig": "distributed/optimized-gpt2-500m--configuration_gpt_optimized.GPTOptimConfig",
268
- "AutoModelForCausalLM": "distributed/optimized-gpt2-500m--modeling_gpt_optimized.GPTOptim"
269
- },
270
  "block_list": [
271
- 5550027,
272
- 5550039,
273
- 5550044,
274
- 5550048,
275
- 5550052
276
  ],
277
- "block_size": 1024,
278
- "bos_token_id": 50256,
279
- "embd_pdrop": 0.1,
280
- "eos_token_id": 50256,
281
  "initializer_range": 0.02,
282
- "inner_step": 3,
283
- "inner_steps": 0,
284
- "last_allreduce_block": 5549491,
285
- "layer_norm_epsilon": 1e-05,
286
- "model_type": "gpt_optimized",
287
- "n_embd": 1280,
288
- "n_head": 32,
289
- "n_inner": null,
290
- "n_layer": 48,
291
- "n_positions": 1024,
292
- "reorder_and_upcast_attn": false,
293
- "resid_pdrop": 0.1,
294
- "scale_attn_by_inverse_layer_idx": false,
295
- "scale_attn_weights": true,
296
- "summary_activation": null,
297
- "summary_first_dropout": 0.1,
298
- "summary_proj_to_labels": true,
299
- "summary_type": "cls_index",
300
- "summary_use_proj": true,
301
  "torch_dtype": "float32",
302
  "transformers_version": "4.39.3",
303
- "use_cache": true,
304
- "vocab_size": 50257
305
  }
 
1
  {
2
+ "_name_or_path": "distributed/llama-1b",
 
3
  "all_reduce_scores": {
4
  "0": "NON_PARTICIPATING",
5
  "1": "SUCCESS",
6
  "10": "NON_PARTICIPATING",
7
  "100": "NON_PARTICIPATING",
8
  "101": "NON_PARTICIPATING",
9
+ "102": "NON_PARTICIPATING",
10
  "103": "NON_PARTICIPATING",
11
  "104": "NON_PARTICIPATING",
12
  "105": "NON_PARTICIPATING",
 
25
  "117": "NON_PARTICIPATING",
26
  "118": "NON_PARTICIPATING",
27
  "119": "NON_PARTICIPATING",
28
+ "12": "NON_PARTICIPATING",
29
  "120": "NON_PARTICIPATING",
30
  "121": "NON_PARTICIPATING",
31
  "122": "NON_PARTICIPATING",
32
  "123": "NON_PARTICIPATING",
33
+ "124": "NON_PARTICIPATING",
34
  "125": "NON_PARTICIPATING",
35
  "126": "NON_PARTICIPATING",
36
  "127": "NON_PARTICIPATING",
37
  "128": "NON_PARTICIPATING",
38
  "129": "NON_PARTICIPATING",
39
  "13": "SUCCESS",
40
+ "130": "NON_PARTICIPATING",
41
  "131": "NON_PARTICIPATING",
42
  "132": "NON_PARTICIPATING",
43
  "133": "NON_PARTICIPATING",
44
+ "134": "SUCCESS",
45
  "135": "NON_PARTICIPATING",
46
  "136": "NON_PARTICIPATING",
47
+ "137": "NON_PARTICIPATING",
48
  "138": "NON_PARTICIPATING",
49
  "139": "NON_PARTICIPATING",
50
  "14": "SUCCESS",
 
55
  "144": "NON_PARTICIPATING",
56
  "145": "NON_PARTICIPATING",
57
  "146": "NON_PARTICIPATING",
58
+ "147": "NON_PARTICIPATING",
59
  "148": "NON_PARTICIPATING",
60
  "149": "NON_PARTICIPATING",
61
  "15": "SUCCESS",
62
+ "150": "NON_PARTICIPATING",
63
  "151": "NON_PARTICIPATING",
64
  "152": "NON_PARTICIPATING",
65
+ "153": "NON_PARTICIPATING",
66
  "154": "NON_PARTICIPATING",
67
  "155": "NON_PARTICIPATING",
68
  "156": "NON_PARTICIPATING",
 
72
  "16": "SUCCESS",
73
  "160": "NON_PARTICIPATING",
74
  "161": "NON_PARTICIPATING",
75
+ "162": "NON_PARTICIPATING",
76
  "163": "NON_PARTICIPATING",
77
  "164": "NON_PARTICIPATING",
78
  "165": "NON_PARTICIPATING",
 
91
  "177": "NON_PARTICIPATING",
92
  "178": "NON_PARTICIPATING",
93
  "179": "NON_PARTICIPATING",
94
+ "18": "NON_PARTICIPATING",
95
  "180": "NON_PARTICIPATING",
96
  "181": "NON_PARTICIPATING",
97
  "182": "NON_PARTICIPATING",
98
  "183": "NON_PARTICIPATING",
99
  "184": "NON_PARTICIPATING",
100
+ "185": "NON_PARTICIPATING",
101
+ "186": "NON_PARTICIPATING",
102
  "187": "NON_PARTICIPATING",
103
  "188": "NON_PARTICIPATING",
104
  "189": "NON_PARTICIPATING",
 
115
  "199": "NON_PARTICIPATING",
116
  "2": "SUCCESS",
117
  "20": "SUCCESS",
118
+ "200": "SUCCESS",
119
  "201": "NON_PARTICIPATING",
120
  "202": "NON_PARTICIPATING",
121
+ "203": "NON_PARTICIPATING",
122
+ "204": "NON_PARTICIPATING",
123
+ "205": "NON_PARTICIPATING",
124
  "206": "NON_PARTICIPATING",
125
  "207": "NON_PARTICIPATING",
126
  "208": "SUCCESS",
 
134
  "215": "NON_PARTICIPATING",
135
  "216": "NON_PARTICIPATING",
136
  "217": "NON_PARTICIPATING",
137
+ "218": "SUCCESS",
138
  "219": "NON_PARTICIPATING",
139
  "22": "SUCCESS",
140
  "220": "NON_PARTICIPATING",
 
146
  "226": "NON_PARTICIPATING",
147
  "227": "NON_PARTICIPATING",
148
  "228": "NON_PARTICIPATING",
149
+ "229": "NON_PARTICIPATING",
150
+ "23": "NON_PARTICIPATING",
151
  "230": "NON_PARTICIPATING",
152
+ "231": "NON_PARTICIPATING",
153
  "232": "NON_PARTICIPATING",
154
  "233": "NON_PARTICIPATING",
155
  "234": "NON_PARTICIPATING",
 
163
  "241": "NON_PARTICIPATING",
164
  "242": "NON_PARTICIPATING",
165
  "243": "NON_PARTICIPATING",
166
+ "244": "NON_PARTICIPATING",
167
  "245": "NON_PARTICIPATING",
168
+ "246": "SUCCESS",
169
  "247": "NON_PARTICIPATING",
170
  "248": "NON_PARTICIPATING",
171
+ "249": "SUCCESS",
172
  "25": "SUCCESS",
173
  "250": "NON_PARTICIPATING",
174
  "251": "NON_PARTICIPATING",
175
+ "252": "NON_PARTICIPATING",
176
  "253": "NON_PARTICIPATING",
177
  "254": "NON_PARTICIPATING",
178
  "255": "NON_PARTICIPATING",
179
+ "26": "NON_PARTICIPATING",
180
  "27": "SUCCESS",
181
  "28": "SUCCESS",
182
  "29": "SUCCESS",
 
185
  "31": "SUCCESS",
186
  "32": "SUCCESS",
187
  "33": "NON_PARTICIPATING",
188
+ "34": "NON_PARTICIPATING",
189
  "35": "SUCCESS",
190
  "36": "SUCCESS",
191
  "37": "SUCCESS",
 
198
  "43": "SUCCESS",
199
  "44": "SUCCESS",
200
  "45": "SUCCESS",
201
+ "46": "SUCCESS",
202
+ "47": "SUCCESS",
203
+ "48": "SUCCESS",
204
  "49": "SUCCESS",
205
  "5": "SUCCESS",
206
+ "50": "NON_PARTICIPATING",
207
  "51": "SUCCESS",
208
  "52": "SUCCESS",
209
  "53": "SUCCESS",
 
211
  "55": "SUCCESS",
212
  "56": "SUCCESS",
213
  "57": "SUCCESS",
214
+ "58": "SUCCESS",
215
  "59": "SUCCESS",
216
+ "6": "SUCCESS",
217
  "60": "SUCCESS",
218
+ "61": "SUCCESS",
219
  "62": "SUCCESS",
220
  "63": "SUCCESS",
221
+ "64": "SUCCESS",
222
  "65": "SUCCESS",
223
  "66": "SUCCESS",
224
  "67": "SUCCESS",
225
  "68": "SUCCESS",
226
  "69": "SUCCESS",
227
  "7": "SUCCESS",
228
+ "70": "SUCCESS",
229
+ "71": "SUCCESS",
230
  "72": "SUCCESS",
231
  "73": "NON_PARTICIPATING",
232
  "74": "SUCCESS",
 
234
  "76": "SUCCESS",
235
  "77": "SUCCESS",
236
  "78": "SUCCESS",
237
+ "79": "SUCCESS",
238
  "8": "SUCCESS",
239
+ "80": "NON_PARTICIPATING",
240
  "81": "SUCCESS",
241
  "82": "SUCCESS",
242
  "83": "NON_PARTICIPATING",
 
244
  "85": "SUCCESS",
245
  "86": "SUCCESS",
246
  "87": "SUCCESS",
247
+ "88": "NON_PARTICIPATING",
248
  "89": "SUCCESS",
249
+ "9": "SUCCESS",
250
+ "90": "NON_PARTICIPATING",
251
  "91": "SUCCESS",
252
  "92": "SUCCESS",
253
+ "93": "SUCCESS",
254
  "94": "SUCCESS",
255
  "95": "SUCCESS",
256
+ "96": "NON_PARTICIPATING",
257
  "97": "SUCCESS",
258
  "98": "SUCCESS",
259
  "99": "SUCCESS"
260
  },
261
  "architectures": [
262
+ "LlamaForCausalLM"
263
  ],
264
+ "attention_bias": false,
265
+ "attention_dropout": 0.0,
 
 
 
266
  "block_list": [
267
+ 5740922,
268
+ 5740952
 
 
 
269
  ],
270
+ "bos_token_id": 1,
271
+ "eos_token_id": 2,
272
+ "hidden_act": "silu",
273
+ "hidden_size": 2048,
274
  "initializer_range": 0.02,
275
+ "inner_step": 2,
276
+ "intermediate_size": 5632,
277
+ "last_allreduce_block": 5740805,
278
+ "max_position_embeddings": 2048,
279
+ "mlp_bias": false,
280
+ "model_type": "llama",
281
+ "num_attention_heads": 32,
282
+ "num_hidden_layers": 22,
283
+ "num_key_value_heads": 4,
284
+ "pretraining_tp": 1,
285
+ "rms_norm_eps": 1e-05,
286
+ "rope_scaling": null,
287
+ "rope_theta": 10000.0,
288
+ "tie_word_embeddings": false,
 
 
 
 
 
289
  "torch_dtype": "float32",
290
  "transformers_version": "4.39.3",
291
+ "use_cache": false,
292
+ "vocab_size": 32000
293
  }
generation_config.json ADDED
@@ -0,0 +1,7 @@
 
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.39.3",
6
+ "use_cache": false
7
+ }
inner_optimizer.pt CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ea1c82a21dcd3724f25af230d0a8b9fb35824fef4bd25445d2daea658b98ef6f
3
- size 8081782026
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5eda789d3d9c4aeaa35631490c431a91385510c2edc48c6bcf1f60a36290c041
3
+ size 8800564076
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:ee53b9498e930243ecea6243de09b64f70dd147c6a8d249d7959b61002ec5a8d
3
- size 4040701744
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:98983797a1411486858b4e49336dc9cc2c1b63f9f28a6400596c21d812683562
3
+ size 4400216536