Text Generation
Transformers
ONNX
llama
sparse
instruct
deepsparse
shubhrapandit commited on
Commit
a076943
1 Parent(s): 297c316

Update model files

Browse files
arc_challenge.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "results": {
3
+ "arc_challenge": {
4
+ "acc": 0.47525597269624575,
5
+ "acc_stderr": 0.014593487694937738,
6
+ "acc_norm": 0.4974402730375427,
7
+ "acc_norm_stderr": 0.014611199329843791
8
+ }
9
+ },
10
+ "versions": {
11
+ "arc_challenge": 0
12
+ },
13
+ "config": {
14
+ "model": "sparseml",
15
+ "model_args": "pretrained=/cache/shubhra/models/platypus_dolphin/cerebras/spft-cerebras_llama2_sparse50_45B_platypus_dolphin_KDFalse_GCTrue_LR1e-4_E2_quant_loge,trust_remote_code=True,dtype=bfloat16",
16
+ "num_fewshot": 25,
17
+ "batch_size": "16",
18
+ "batch_sizes": [],
19
+ "device": "cuda:0",
20
+ "no_cache": true,
21
+ "limit": null,
22
+ "bootstrap_iters": 100000,
23
+ "description_dict": {}
24
+ }
25
+ }
config.json CHANGED
@@ -1,4 +1,5 @@
1
  {
 
2
  "architectures": [
3
  "LlamaForCausalLM"
4
  ],
@@ -22,7 +23,7 @@
22
  "tie_word_embeddings": false,
23
  "tokenizer_class": "LlamaTokenizerFast",
24
  "torch_dtype": "float32",
25
- "transformers_version": "1.7.0.43401",
26
  "use_cache": true,
27
  "vocab_size": 32000
28
  }
 
1
  {
2
+ "_name_or_path": "/cache/shubhra/models/platypus_dolphin/cerebras/spft-cerebras_llama2_sparse50_45B_platypus_dolphin_KDFalse_GCTrue_LR1e-4_E2/combined/",
3
  "architectures": [
4
  "LlamaForCausalLM"
5
  ],
 
23
  "tie_word_embeddings": false,
24
  "tokenizer_class": "LlamaTokenizerFast",
25
  "torch_dtype": "float32",
26
+ "transformers_version": "4.39.3",
27
  "use_cache": true,
28
  "vocab_size": 32000
29
  }
generation_config.json ADDED
@@ -0,0 +1,6 @@
 
 
 
 
 
 
 
1
+ {
2
+ "_from_model_config": true,
3
+ "bos_token_id": 1,
4
+ "eos_token_id": 2,
5
+ "transformers_version": "4.39.3"
6
+ }
gsm8k.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "results": {
3
+ "gsm8k": {
4
+ "acc": 0.17589082638362397,
5
+ "acc_stderr": 0.010487120635539611
6
+ }
7
+ },
8
+ "versions": {
9
+ "gsm8k": 0
10
+ },
11
+ "config": {
12
+ "model": "sparseml",
13
+ "model_args": "pretrained=/cache/shubhra/models/platypus_dolphin/cerebras/spft-cerebras_llama2_sparse50_45B_platypus_dolphin_KDFalse_GCTrue_LR1e-4_E2_quant_loge,trust_remote_code=True,dtype=bfloat16",
14
+ "num_fewshot": 5,
15
+ "batch_size": "16",
16
+ "batch_sizes": [],
17
+ "device": "cuda:0",
18
+ "no_cache": true,
19
+ "limit": null,
20
+ "bootstrap_iters": 100000,
21
+ "description_dict": {}
22
+ }
23
+ }
hellaswag.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "results": {
3
+ "hellaswag": {
4
+ "acc": 0.5891256721768572,
5
+ "acc_stderr": 0.004909870006388837,
6
+ "acc_norm": 0.7780322644891456,
7
+ "acc_norm_stderr": 0.004147202539759584
8
+ }
9
+ },
10
+ "versions": {
11
+ "hellaswag": 0
12
+ },
13
+ "config": {
14
+ "model": "sparseml",
15
+ "model_args": "pretrained=/cache/shubhra/models/platypus_dolphin/cerebras/spft-cerebras_llama2_sparse50_45B_platypus_dolphin_KDFalse_GCTrue_LR1e-4_E2_quant_loge,trust_remote_code=True,dtype=bfloat16",
16
+ "num_fewshot": 10,
17
+ "batch_size": "16",
18
+ "batch_sizes": [],
19
+ "device": "cuda:0",
20
+ "no_cache": true,
21
+ "limit": null,
22
+ "bootstrap_iters": 100000,
23
+ "description_dict": {}
24
+ }
25
+ }
mmlu.json ADDED
@@ -0,0 +1,417 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "results": {
3
+ "hendrycksTest-abstract_algebra": {
4
+ "acc": 0.28,
5
+ "acc_stderr": 0.04512608598542128,
6
+ "acc_norm": 0.28,
7
+ "acc_norm_stderr": 0.04512608598542128
8
+ },
9
+ "hendrycksTest-anatomy": {
10
+ "acc": 0.4740740740740741,
11
+ "acc_stderr": 0.04313531696750574,
12
+ "acc_norm": 0.4740740740740741,
13
+ "acc_norm_stderr": 0.04313531696750574
14
+ },
15
+ "hendrycksTest-astronomy": {
16
+ "acc": 0.5131578947368421,
17
+ "acc_stderr": 0.04067533136309174,
18
+ "acc_norm": 0.5131578947368421,
19
+ "acc_norm_stderr": 0.04067533136309174
20
+ },
21
+ "hendrycksTest-business_ethics": {
22
+ "acc": 0.49,
23
+ "acc_stderr": 0.05024183937956912,
24
+ "acc_norm": 0.49,
25
+ "acc_norm_stderr": 0.05024183937956912
26
+ },
27
+ "hendrycksTest-clinical_knowledge": {
28
+ "acc": 0.4867924528301887,
29
+ "acc_stderr": 0.030762134874500482,
30
+ "acc_norm": 0.4867924528301887,
31
+ "acc_norm_stderr": 0.030762134874500482
32
+ },
33
+ "hendrycksTest-college_biology": {
34
+ "acc": 0.4652777777777778,
35
+ "acc_stderr": 0.04171115858181618,
36
+ "acc_norm": 0.4652777777777778,
37
+ "acc_norm_stderr": 0.04171115858181618
38
+ },
39
+ "hendrycksTest-college_chemistry": {
40
+ "acc": 0.32,
41
+ "acc_stderr": 0.046882617226215034,
42
+ "acc_norm": 0.32,
43
+ "acc_norm_stderr": 0.046882617226215034
44
+ },
45
+ "hendrycksTest-college_computer_science": {
46
+ "acc": 0.36,
47
+ "acc_stderr": 0.04824181513244218,
48
+ "acc_norm": 0.36,
49
+ "acc_norm_stderr": 0.04824181513244218
50
+ },
51
+ "hendrycksTest-college_mathematics": {
52
+ "acc": 0.26,
53
+ "acc_stderr": 0.04408440022768078,
54
+ "acc_norm": 0.26,
55
+ "acc_norm_stderr": 0.04408440022768078
56
+ },
57
+ "hendrycksTest-college_medicine": {
58
+ "acc": 0.37572254335260113,
59
+ "acc_stderr": 0.036928207672648664,
60
+ "acc_norm": 0.37572254335260113,
61
+ "acc_norm_stderr": 0.036928207672648664
62
+ },
63
+ "hendrycksTest-college_physics": {
64
+ "acc": 0.24509803921568626,
65
+ "acc_stderr": 0.04280105837364395,
66
+ "acc_norm": 0.24509803921568626,
67
+ "acc_norm_stderr": 0.04280105837364395
68
+ },
69
+ "hendrycksTest-computer_security": {
70
+ "acc": 0.56,
71
+ "acc_stderr": 0.04988876515698589,
72
+ "acc_norm": 0.56,
73
+ "acc_norm_stderr": 0.04988876515698589
74
+ },
75
+ "hendrycksTest-conceptual_physics": {
76
+ "acc": 0.3872340425531915,
77
+ "acc_stderr": 0.03184389265339526,
78
+ "acc_norm": 0.3872340425531915,
79
+ "acc_norm_stderr": 0.03184389265339526
80
+ },
81
+ "hendrycksTest-econometrics": {
82
+ "acc": 0.2719298245614035,
83
+ "acc_stderr": 0.04185774424022056,
84
+ "acc_norm": 0.2719298245614035,
85
+ "acc_norm_stderr": 0.04185774424022056
86
+ },
87
+ "hendrycksTest-electrical_engineering": {
88
+ "acc": 0.4206896551724138,
89
+ "acc_stderr": 0.0411391498118926,
90
+ "acc_norm": 0.4206896551724138,
91
+ "acc_norm_stderr": 0.0411391498118926
92
+ },
93
+ "hendrycksTest-elementary_mathematics": {
94
+ "acc": 0.30687830687830686,
95
+ "acc_stderr": 0.023752928712112143,
96
+ "acc_norm": 0.30687830687830686,
97
+ "acc_norm_stderr": 0.023752928712112143
98
+ },
99
+ "hendrycksTest-formal_logic": {
100
+ "acc": 0.2698412698412698,
101
+ "acc_stderr": 0.03970158273235173,
102
+ "acc_norm": 0.2698412698412698,
103
+ "acc_norm_stderr": 0.03970158273235173
104
+ },
105
+ "hendrycksTest-global_facts": {
106
+ "acc": 0.28,
107
+ "acc_stderr": 0.04512608598542127,
108
+ "acc_norm": 0.28,
109
+ "acc_norm_stderr": 0.04512608598542127
110
+ },
111
+ "hendrycksTest-high_school_biology": {
112
+ "acc": 0.5225806451612903,
113
+ "acc_stderr": 0.02841498501970786,
114
+ "acc_norm": 0.5225806451612903,
115
+ "acc_norm_stderr": 0.02841498501970786
116
+ },
117
+ "hendrycksTest-high_school_chemistry": {
118
+ "acc": 0.3054187192118227,
119
+ "acc_stderr": 0.03240661565868408,
120
+ "acc_norm": 0.3054187192118227,
121
+ "acc_norm_stderr": 0.03240661565868408
122
+ },
123
+ "hendrycksTest-high_school_computer_science": {
124
+ "acc": 0.43,
125
+ "acc_stderr": 0.049756985195624284,
126
+ "acc_norm": 0.43,
127
+ "acc_norm_stderr": 0.049756985195624284
128
+ },
129
+ "hendrycksTest-high_school_european_history": {
130
+ "acc": 0.5636363636363636,
131
+ "acc_stderr": 0.03872592983524754,
132
+ "acc_norm": 0.5636363636363636,
133
+ "acc_norm_stderr": 0.03872592983524754
134
+ },
135
+ "hendrycksTest-high_school_geography": {
136
+ "acc": 0.48484848484848486,
137
+ "acc_stderr": 0.03560716516531061,
138
+ "acc_norm": 0.48484848484848486,
139
+ "acc_norm_stderr": 0.03560716516531061
140
+ },
141
+ "hendrycksTest-high_school_government_and_politics": {
142
+ "acc": 0.6373056994818653,
143
+ "acc_stderr": 0.03469713791704371,
144
+ "acc_norm": 0.6373056994818653,
145
+ "acc_norm_stderr": 0.03469713791704371
146
+ },
147
+ "hendrycksTest-high_school_macroeconomics": {
148
+ "acc": 0.4282051282051282,
149
+ "acc_stderr": 0.02508830145469483,
150
+ "acc_norm": 0.4282051282051282,
151
+ "acc_norm_stderr": 0.02508830145469483
152
+ },
153
+ "hendrycksTest-high_school_mathematics": {
154
+ "acc": 0.26296296296296295,
155
+ "acc_stderr": 0.026842057873833706,
156
+ "acc_norm": 0.26296296296296295,
157
+ "acc_norm_stderr": 0.026842057873833706
158
+ },
159
+ "hendrycksTest-high_school_microeconomics": {
160
+ "acc": 0.40756302521008403,
161
+ "acc_stderr": 0.03191863374478465,
162
+ "acc_norm": 0.40756302521008403,
163
+ "acc_norm_stderr": 0.03191863374478465
164
+ },
165
+ "hendrycksTest-high_school_physics": {
166
+ "acc": 0.2781456953642384,
167
+ "acc_stderr": 0.03658603262763744,
168
+ "acc_norm": 0.2781456953642384,
169
+ "acc_norm_stderr": 0.03658603262763744
170
+ },
171
+ "hendrycksTest-high_school_psychology": {
172
+ "acc": 0.5559633027522936,
173
+ "acc_stderr": 0.02130262121165452,
174
+ "acc_norm": 0.5559633027522936,
175
+ "acc_norm_stderr": 0.02130262121165452
176
+ },
177
+ "hendrycksTest-high_school_statistics": {
178
+ "acc": 0.37037037037037035,
179
+ "acc_stderr": 0.03293377139415191,
180
+ "acc_norm": 0.37037037037037035,
181
+ "acc_norm_stderr": 0.03293377139415191
182
+ },
183
+ "hendrycksTest-high_school_us_history": {
184
+ "acc": 0.6029411764705882,
185
+ "acc_stderr": 0.0343413116471913,
186
+ "acc_norm": 0.6029411764705882,
187
+ "acc_norm_stderr": 0.0343413116471913
188
+ },
189
+ "hendrycksTest-high_school_world_history": {
190
+ "acc": 0.6582278481012658,
191
+ "acc_stderr": 0.03087453753755362,
192
+ "acc_norm": 0.6582278481012658,
193
+ "acc_norm_stderr": 0.03087453753755362
194
+ },
195
+ "hendrycksTest-human_aging": {
196
+ "acc": 0.5515695067264574,
197
+ "acc_stderr": 0.033378837362550984,
198
+ "acc_norm": 0.5515695067264574,
199
+ "acc_norm_stderr": 0.033378837362550984
200
+ },
201
+ "hendrycksTest-human_sexuality": {
202
+ "acc": 0.5343511450381679,
203
+ "acc_stderr": 0.04374928560599738,
204
+ "acc_norm": 0.5343511450381679,
205
+ "acc_norm_stderr": 0.04374928560599738
206
+ },
207
+ "hendrycksTest-international_law": {
208
+ "acc": 0.6776859504132231,
209
+ "acc_stderr": 0.04266416363352167,
210
+ "acc_norm": 0.6776859504132231,
211
+ "acc_norm_stderr": 0.04266416363352167
212
+ },
213
+ "hendrycksTest-jurisprudence": {
214
+ "acc": 0.5,
215
+ "acc_stderr": 0.04833682445228318,
216
+ "acc_norm": 0.5,
217
+ "acc_norm_stderr": 0.04833682445228318
218
+ },
219
+ "hendrycksTest-logical_fallacies": {
220
+ "acc": 0.4601226993865031,
221
+ "acc_stderr": 0.03915857291436972,
222
+ "acc_norm": 0.4601226993865031,
223
+ "acc_norm_stderr": 0.03915857291436972
224
+ },
225
+ "hendrycksTest-machine_learning": {
226
+ "acc": 0.30357142857142855,
227
+ "acc_stderr": 0.04364226155841044,
228
+ "acc_norm": 0.30357142857142855,
229
+ "acc_norm_stderr": 0.04364226155841044
230
+ },
231
+ "hendrycksTest-management": {
232
+ "acc": 0.5922330097087378,
233
+ "acc_stderr": 0.048657775704107696,
234
+ "acc_norm": 0.5922330097087378,
235
+ "acc_norm_stderr": 0.048657775704107696
236
+ },
237
+ "hendrycksTest-marketing": {
238
+ "acc": 0.6324786324786325,
239
+ "acc_stderr": 0.031585391577456365,
240
+ "acc_norm": 0.6324786324786325,
241
+ "acc_norm_stderr": 0.031585391577456365
242
+ },
243
+ "hendrycksTest-medical_genetics": {
244
+ "acc": 0.45,
245
+ "acc_stderr": 0.05,
246
+ "acc_norm": 0.45,
247
+ "acc_norm_stderr": 0.05
248
+ },
249
+ "hendrycksTest-miscellaneous": {
250
+ "acc": 0.632183908045977,
251
+ "acc_stderr": 0.017243828891846263,
252
+ "acc_norm": 0.632183908045977,
253
+ "acc_norm_stderr": 0.017243828891846263
254
+ },
255
+ "hendrycksTest-moral_disputes": {
256
+ "acc": 0.49710982658959535,
257
+ "acc_stderr": 0.026918645383239015,
258
+ "acc_norm": 0.49710982658959535,
259
+ "acc_norm_stderr": 0.026918645383239015
260
+ },
261
+ "hendrycksTest-moral_scenarios": {
262
+ "acc": 0.2748603351955307,
263
+ "acc_stderr": 0.01493131670322051,
264
+ "acc_norm": 0.2748603351955307,
265
+ "acc_norm_stderr": 0.01493131670322051
266
+ },
267
+ "hendrycksTest-nutrition": {
268
+ "acc": 0.477124183006536,
269
+ "acc_stderr": 0.028599936776089782,
270
+ "acc_norm": 0.477124183006536,
271
+ "acc_norm_stderr": 0.028599936776089782
272
+ },
273
+ "hendrycksTest-philosophy": {
274
+ "acc": 0.5562700964630225,
275
+ "acc_stderr": 0.028217683556652315,
276
+ "acc_norm": 0.5562700964630225,
277
+ "acc_norm_stderr": 0.028217683556652315
278
+ },
279
+ "hendrycksTest-prehistory": {
280
+ "acc": 0.5123456790123457,
281
+ "acc_stderr": 0.027812262269327242,
282
+ "acc_norm": 0.5123456790123457,
283
+ "acc_norm_stderr": 0.027812262269327242
284
+ },
285
+ "hendrycksTest-professional_accounting": {
286
+ "acc": 0.35106382978723405,
287
+ "acc_stderr": 0.028473501272963764,
288
+ "acc_norm": 0.35106382978723405,
289
+ "acc_norm_stderr": 0.028473501272963764
290
+ },
291
+ "hendrycksTest-professional_law": {
292
+ "acc": 0.3683181225554107,
293
+ "acc_stderr": 0.012319403369564639,
294
+ "acc_norm": 0.3683181225554107,
295
+ "acc_norm_stderr": 0.012319403369564639
296
+ },
297
+ "hendrycksTest-professional_medicine": {
298
+ "acc": 0.4264705882352941,
299
+ "acc_stderr": 0.030042615832714867,
300
+ "acc_norm": 0.4264705882352941,
301
+ "acc_norm_stderr": 0.030042615832714867
302
+ },
303
+ "hendrycksTest-professional_psychology": {
304
+ "acc": 0.4411764705882353,
305
+ "acc_stderr": 0.020087362076702853,
306
+ "acc_norm": 0.4411764705882353,
307
+ "acc_norm_stderr": 0.020087362076702853
308
+ },
309
+ "hendrycksTest-public_relations": {
310
+ "acc": 0.5272727272727272,
311
+ "acc_stderr": 0.04782001791380061,
312
+ "acc_norm": 0.5272727272727272,
313
+ "acc_norm_stderr": 0.04782001791380061
314
+ },
315
+ "hendrycksTest-security_studies": {
316
+ "acc": 0.5265306122448979,
317
+ "acc_stderr": 0.03196412734523272,
318
+ "acc_norm": 0.5265306122448979,
319
+ "acc_norm_stderr": 0.03196412734523272
320
+ },
321
+ "hendrycksTest-sociology": {
322
+ "acc": 0.582089552238806,
323
+ "acc_stderr": 0.034875586404620636,
324
+ "acc_norm": 0.582089552238806,
325
+ "acc_norm_stderr": 0.034875586404620636
326
+ },
327
+ "hendrycksTest-us_foreign_policy": {
328
+ "acc": 0.71,
329
+ "acc_stderr": 0.045604802157206845,
330
+ "acc_norm": 0.71,
331
+ "acc_norm_stderr": 0.045604802157206845
332
+ },
333
+ "hendrycksTest-virology": {
334
+ "acc": 0.41566265060240964,
335
+ "acc_stderr": 0.03836722176598052,
336
+ "acc_norm": 0.41566265060240964,
337
+ "acc_norm_stderr": 0.03836722176598052
338
+ },
339
+ "hendrycksTest-world_religions": {
340
+ "acc": 0.6783625730994152,
341
+ "acc_stderr": 0.03582529442573122,
342
+ "acc_norm": 0.6783625730994152,
343
+ "acc_norm_stderr": 0.03582529442573122
344
+ }
345
+ },
346
+ "versions": {
347
+ "hendrycksTest-abstract_algebra": 1,
348
+ "hendrycksTest-anatomy": 1,
349
+ "hendrycksTest-astronomy": 1,
350
+ "hendrycksTest-business_ethics": 1,
351
+ "hendrycksTest-clinical_knowledge": 1,
352
+ "hendrycksTest-college_biology": 1,
353
+ "hendrycksTest-college_chemistry": 1,
354
+ "hendrycksTest-college_computer_science": 1,
355
+ "hendrycksTest-college_mathematics": 1,
356
+ "hendrycksTest-college_medicine": 1,
357
+ "hendrycksTest-college_physics": 1,
358
+ "hendrycksTest-computer_security": 1,
359
+ "hendrycksTest-conceptual_physics": 1,
360
+ "hendrycksTest-econometrics": 1,
361
+ "hendrycksTest-electrical_engineering": 1,
362
+ "hendrycksTest-elementary_mathematics": 1,
363
+ "hendrycksTest-formal_logic": 1,
364
+ "hendrycksTest-global_facts": 1,
365
+ "hendrycksTest-high_school_biology": 1,
366
+ "hendrycksTest-high_school_chemistry": 1,
367
+ "hendrycksTest-high_school_computer_science": 1,
368
+ "hendrycksTest-high_school_european_history": 1,
369
+ "hendrycksTest-high_school_geography": 1,
370
+ "hendrycksTest-high_school_government_and_politics": 1,
371
+ "hendrycksTest-high_school_macroeconomics": 1,
372
+ "hendrycksTest-high_school_mathematics": 1,
373
+ "hendrycksTest-high_school_microeconomics": 1,
374
+ "hendrycksTest-high_school_physics": 1,
375
+ "hendrycksTest-high_school_psychology": 1,
376
+ "hendrycksTest-high_school_statistics": 1,
377
+ "hendrycksTest-high_school_us_history": 1,
378
+ "hendrycksTest-high_school_world_history": 1,
379
+ "hendrycksTest-human_aging": 1,
380
+ "hendrycksTest-human_sexuality": 1,
381
+ "hendrycksTest-international_law": 1,
382
+ "hendrycksTest-jurisprudence": 1,
383
+ "hendrycksTest-logical_fallacies": 1,
384
+ "hendrycksTest-machine_learning": 1,
385
+ "hendrycksTest-management": 1,
386
+ "hendrycksTest-marketing": 1,
387
+ "hendrycksTest-medical_genetics": 1,
388
+ "hendrycksTest-miscellaneous": 1,
389
+ "hendrycksTest-moral_disputes": 1,
390
+ "hendrycksTest-moral_scenarios": 1,
391
+ "hendrycksTest-nutrition": 1,
392
+ "hendrycksTest-philosophy": 1,
393
+ "hendrycksTest-prehistory": 1,
394
+ "hendrycksTest-professional_accounting": 1,
395
+ "hendrycksTest-professional_law": 1,
396
+ "hendrycksTest-professional_medicine": 1,
397
+ "hendrycksTest-professional_psychology": 1,
398
+ "hendrycksTest-public_relations": 1,
399
+ "hendrycksTest-security_studies": 1,
400
+ "hendrycksTest-sociology": 1,
401
+ "hendrycksTest-us_foreign_policy": 1,
402
+ "hendrycksTest-virology": 1,
403
+ "hendrycksTest-world_religions": 1
404
+ },
405
+ "config": {
406
+ "model": "sparseml",
407
+ "model_args": "pretrained=/cache/shubhra/models/platypus_dolphin/cerebras/spft-cerebras_llama2_sparse50_45B_platypus_dolphin_KDFalse_GCTrue_LR1e-4_E2_quant_loge,trust_remote_code=True,dtype=bfloat16",
408
+ "num_fewshot": 5,
409
+ "batch_size": "8",
410
+ "batch_sizes": [],
411
+ "device": "cuda:0",
412
+ "no_cache": true,
413
+ "limit": null,
414
+ "bootstrap_iters": 100000,
415
+ "description_dict": {}
416
+ }
417
+ }
model-orig.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:9ee7a53f34780e582d10cf2a9403b7135f822cd3c2589756d030d081a775bee4
3
- size 1048493
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f48e7f4d7d6ef80235433b0103c086af02b340596531b2e092ba8c0fae8e89d0
3
+ size 1047345
model.data CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:1430185bc4cc7793e682e4b0f817eff28e1d1ea61199b9f70705a65f447497ce
3
- size 7290022912
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5e76f571fe1648c66705dfdc0df539fb5ce9b864f54051df301fadea45986056
3
+ size 7425272832
model.onnx CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:56e2f3c2001e0afcf804e11766e0afb6c7d8227177b0b7c223c63a855ab05291
3
- size 1033438
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d288114f93740b56b7062217b4735aa8c4ded2e9393b92e84fe78064112dfe65
3
+ size 1032290
recipe.yaml CHANGED
@@ -11,8 +11,8 @@ test_stage:
11
  QuantizationModifier:
12
  ignore: [LlamaRotaryEmbedding, LlamaRMSNorm, SiLUActivation, model.layers.30.mlp.down_proj,
13
  model.layers.1.mlp.down_proj, model.layers.0.mlp.down_proj, model.layers.4.mlp.down_proj,
14
- MatMulOutput_QK, MatMulOutput_PV, MatMulLeftInput_QK, MatMulLeftInput_PV, MatMulRightInput_QK,
15
- MatMulRightInput_PV, QuantizableMatMul]
16
  post_oneshot_calibration: true
17
  scheme_overrides:
18
  Linear:
@@ -34,5 +34,5 @@ test_stage:
34
  model.layers.16, model.layers.17, model.layers.18, model.layers.19, model.layers.20,
35
  model.layers.21, model.layers.22, model.layers.23, model.layers.24, model.layers.25,
36
  model.layers.26, model.layers.27, model.layers.28, model.layers.29, model.layers.30,
37
- model.layers.31]
38
  target_ids: [attention_mask, position_ids]
 
11
  QuantizationModifier:
12
  ignore: [LlamaRotaryEmbedding, LlamaRMSNorm, SiLUActivation, model.layers.30.mlp.down_proj,
13
  model.layers.1.mlp.down_proj, model.layers.0.mlp.down_proj, model.layers.4.mlp.down_proj,
14
+ model.layers.8.mlp.down_proj, MatMulOutput_QK, MatMulOutput_PV, MatMulLeftInput_QK,
15
+ MatMulLeftInput_PV, MatMulRightInput_QK, MatMulRightInput_PV, QuantizableMatMul]
16
  post_oneshot_calibration: true
17
  scheme_overrides:
18
  Linear:
 
34
  model.layers.16, model.layers.17, model.layers.18, model.layers.19, model.layers.20,
35
  model.layers.21, model.layers.22, model.layers.23, model.layers.24, model.layers.25,
36
  model.layers.26, model.layers.27, model.layers.28, model.layers.29, model.layers.30,
37
+ model.layers.31, lm_head]
38
  target_ids: [attention_mask, position_ids]
tokenizer.json CHANGED
@@ -1,11 +1,6 @@
1
  {
2
  "version": "1.0",
3
- "truncation": {
4
- "direction": "Right",
5
- "max_length": 4096,
6
- "strategy": "LongestFirst",
7
- "stride": 0
8
- },
9
  "padding": null,
10
  "added_tokens": [
11
  {
 
1
  {
2
  "version": "1.0",
3
+ "truncation": null,
 
 
 
 
 
4
  "padding": null,
5
  "added_tokens": [
6
  {
tokenizer_config.json CHANGED
@@ -1,4 +1,6 @@
1
  {
 
 
2
  "added_tokens_decoder": {
3
  "0": {
4
  "content": "<unk>",
 
1
  {
2
+ "add_bos_token": true,
3
+ "add_eos_token": false,
4
  "added_tokens_decoder": {
5
  "0": {
6
  "content": "<unk>",
truthfulqa_mc.json ADDED
@@ -0,0 +1,25 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "results": {
3
+ "truthfulqa_mc": {
4
+ "mc1": 0.2741738066095471,
5
+ "mc1_stderr": 0.01561651849721937,
6
+ "mc2": 0.4218222710711044,
7
+ "mc2_stderr": 0.014710008395754494
8
+ }
9
+ },
10
+ "versions": {
11
+ "truthfulqa_mc": 1
12
+ },
13
+ "config": {
14
+ "model": "sparseml",
15
+ "model_args": "pretrained=/cache/shubhra/models/platypus_dolphin/cerebras/spft-cerebras_llama2_sparse50_45B_platypus_dolphin_KDFalse_GCTrue_LR1e-4_E2_quant_loge,trust_remote_code=True,dtype=bfloat16",
16
+ "num_fewshot": 0,
17
+ "batch_size": "16",
18
+ "batch_sizes": [],
19
+ "device": "cuda:0",
20
+ "no_cache": true,
21
+ "limit": null,
22
+ "bootstrap_iters": 100000,
23
+ "description_dict": {}
24
+ }
25
+ }
winogrande.json ADDED
@@ -0,0 +1,23 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "results": {
3
+ "winogrande": {
4
+ "acc": 0.7134964483030781,
5
+ "acc_stderr": 0.012707030139960378
6
+ }
7
+ },
8
+ "versions": {
9
+ "winogrande": 0
10
+ },
11
+ "config": {
12
+ "model": "sparseml",
13
+ "model_args": "pretrained=/cache/shubhra/models/platypus_dolphin/cerebras/spft-cerebras_llama2_sparse50_45B_platypus_dolphin_KDFalse_GCTrue_LR1e-4_E2_quant_loge,trust_remote_code=True,dtype=bfloat16",
14
+ "num_fewshot": 5,
15
+ "batch_size": "16",
16
+ "batch_sizes": [],
17
+ "device": "cuda:0",
18
+ "no_cache": true,
19
+ "limit": null,
20
+ "bootstrap_iters": 100000,
21
+ "description_dict": {}
22
+ }
23
+ }