Hugging Face's logo

tangledgroup
/
tangled-llama-a-128k-base-v0.1

Text Generation
Transformers
Safetensors
llama
litgpt
litdata
conversational
text-generation-inference
Inference Endpoints
Model card Files Community
mtasic85 commited on
Commit
d7d2c9d
•
1 Parent(s): fa41d5c

pretrain eval

Browse files
Files changed (2) hide show
  1. README.md +183 -5
  2. scripts/pretrain-model.yaml +8 -8
README.md CHANGED
@@ -58,14 +58,152 @@ The objective is to streamline the cognitive or reasoning core, eliminating any
58
  litgpt evaluate --tasks 'hellaswag,gsm8k,truthfulqa_mc2,mmlu,winogrande,arc_challenge' --out_dir 'evaluate-quick/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
59
  ```
60
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
61
  ```bash
62
  litgpt evaluate --tasks 'leaderboard' --out_dir 'evaluate-leaderboard/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
63
  ```
64
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
65
  ```bash
66
- litgpt evaluate --tasks 'bbh_zeroshot,bbh_fewshot,bbh_cot_fewshot,bbh_cot_zeroshot' --out_dir 'evaluate-bigbenchhard/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
67
  ```
68
 
 
 
 
 
 
 
 
69
  ```bash
70
  litgpt evaluate --tasks 'mmlu,mmlu_pro' --out_dir 'evaluate-mmlu/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
71
  ```
@@ -74,12 +212,52 @@ litgpt evaluate --tasks 'mmlu,mmlu_pro' --out_dir 'evaluate-mmlu/' --batch_size
74
  litgpt evaluate --tasks 'arc_challenge,boolq,gpqa,hellaswag,openbookqa,piqa,truthfulqa_mc2,winogrande' --out_dir 'evaluate-reasoning/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
75
  ```
76
 
77
- ```bash
78
- litgpt evaluate --tasks 'mmlu_multilingual,mgsm' --out_dir 'evaluate-multilinguals/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
79
- ```
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
80
 
81
  ```bash
82
- litgpt evaluate --tasks 'gsm8k,mathqa' --out_dir 'evaluate-math/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
83
  ```
84
 
85
  ```bash
 
58
  litgpt evaluate --tasks 'hellaswag,gsm8k,truthfulqa_mc2,mmlu,winogrande,arc_challenge' --out_dir 'evaluate-quick/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
59
  ```
60
 
61
+ | Tasks |Version| Filter |n-shot| Metric | |Value | |Stderr|
62
+ |---------------------------------------|------:|----------------|-----:|-----------|---|-----:|---|-----:|
63
+ |arc_challenge | 1|none | 0|acc |↑ |0.2176|± |0.0121|
64
+ | | |none | 0|acc_norm |↑ |0.2560|± |0.0128|
65
+ |gsm8k | 3|flexible-extract| 5|exact_match|↑ |0.0190|± |0.0038|
66
+ | | |strict-match | 5|exact_match|↑ |0.0000|± |0.0000|
67
+ |hellaswag | 1|none | 0|acc |↑ |0.2618|± |0.0044|
68
+ | | |none | 0|acc_norm |↑ |0.2592|± |0.0044|
69
+ |mmlu | 2|none | |acc |↑ |0.2464|± |0.0036|
70
+ | - humanities | 2|none | |acc |↑ |0.2485|± |0.0063|
71
+ | - formal_logic | 1|none | 0|acc |↑ |0.3175|± |0.0416|
72
+ | - high_school_european_history | 1|none | 0|acc |↑ |0.2364|± |0.0332|
73
+ | - high_school_us_history | 1|none | 0|acc |↑ |0.2402|± |0.0300|
74
+ | - high_school_world_history | 1|none | 0|acc |↑ |0.2785|± |0.0292|
75
+ | - international_law | 1|none | 0|acc |↑ |0.2314|± |0.0385|
76
+ | - jurisprudence | 1|none | 0|acc |↑ |0.2407|± |0.0413|
77
+ | - logical_fallacies | 1|none | 0|acc |↑ |0.2086|± |0.0319|
78
+ | - moral_disputes | 1|none | 0|acc |↑ |0.2081|± |0.0219|
79
+ | - moral_scenarios | 1|none | 0|acc |↑ |0.2693|± |0.0148|
80
+ | - philosophy | 1|none | 0|acc |↑ |0.1961|± |0.0226|
81
+ | - prehistory | 1|none | 0|acc |↑ |0.2284|± |0.0234|
82
+ | - professional_law | 1|none | 0|acc |↑ |0.2529|± |0.0111|
83
+ | - world_religions | 1|none | 0|acc |↑ |0.2982|± |0.0351|
84
+ | - other | 2|none | |acc |↑ |0.2536|± |0.0078|
85
+ | - business_ethics | 1|none | 0|acc |↑ |0.2700|± |0.0446|
86
+ | - clinical_knowledge | 1|none | 0|acc |↑ |0.2264|± |0.0258|
87
+ | - college_medicine | 1|none | 0|acc |↑ |0.2312|± |0.0321|
88
+ | - global_facts | 1|none | 0|acc |↑ |0.1500|± |0.0359|
89
+ | - human_aging | 1|none | 0|acc |↑ |0.2242|± |0.0280|
90
+ | - management | 1|none | 0|acc |↑ |0.1942|± |0.0392|
91
+ | - marketing | 1|none | 0|acc |↑ |0.3034|± |0.0301|
92
+ | - medical_genetics | 1|none | 0|acc |↑ |0.2200|± |0.0416|
93
+ | - miscellaneous | 1|none | 0|acc |↑ |0.2401|± |0.0153|
94
+ | - nutrition | 1|none | 0|acc |↑ |0.2255|± |0.0239|
95
+ | - professional_accounting | 1|none | 0|acc |↑ |0.2730|± |0.0266|
96
+ | - professional_medicine | 1|none | 0|acc |↑ |0.4081|± |0.0299|
97
+ | - virology | 1|none | 0|acc |↑ |0.2289|± |0.0327|
98
+ | - social sciences | 2|none | |acc |↑ |0.2535|± |0.0079|
99
+ | - econometrics | 1|none | 0|acc |↑ |0.2368|± |0.0400|
100
+ | - high_school_geography | 1|none | 0|acc |↑ |0.2323|± |0.0301|
101
+ | - high_school_government_and_politics| 1|none | 0|acc |↑ |0.2539|± |0.0314|
102
+ | - high_school_macroeconomics | 1|none | 0|acc |↑ |0.2436|± |0.0218|
103
+ | - high_school_microeconomics | 1|none | 0|acc |↑ |0.2311|± |0.0274|
104
+ | - high_school_psychology | 1|none | 0|acc |↑ |0.2550|± |0.0187|
105
+ | - human_sexuality | 1|none | 0|acc |↑ |0.2824|± |0.0395|
106
+ | - professional_psychology | 1|none | 0|acc |↑ |0.2484|± |0.0175|
107
+ | - public_relations | 1|none | 0|acc |↑ |0.2727|± |0.0427|
108
+ | - security_studies | 1|none | 0|acc |↑ |0.2939|± |0.0292|
109
+ | - sociology | 1|none | 0|acc |↑ |0.2488|± |0.0306|
110
+ | - us_foreign_policy | 1|none | 0|acc |↑ |0.2800|± |0.0451|
111
+ | - stem | 2|none | |acc |↑ |0.2293|± |0.0075|
112
+ | - abstract_algebra | 1|none | 0|acc |↑ |0.2200|± |0.0416|
113
+ | - anatomy | 1|none | 0|acc |↑ |0.2519|± |0.0375|
114
+ | - astronomy | 1|none | 0|acc |↑ |0.2697|± |0.0361|
115
+ | - college_biology | 1|none | 0|acc |↑ |0.2500|± |0.0362|
116
+ | - college_chemistry | 1|none | 0|acc |↑ |0.2400|± |0.0429|
117
+ | - college_computer_science | 1|none | 0|acc |↑ |0.2800|± |0.0451|
118
+ | - college_mathematics | 1|none | 0|acc |↑ |0.2000|± |0.0402|
119
+ | - college_physics | 1|none | 0|acc |↑ |0.2647|± |0.0439|
120
+ | - computer_security | 1|none | 0|acc |↑ |0.1900|± |0.0394|
121
+ | - conceptual_physics | 1|none | 0|acc |↑ |0.2340|± |0.0277|
122
+ | - electrical_engineering | 1|none | 0|acc |↑ |0.2414|± |0.0357|
123
+ | - elementary_mathematics | 1|none | 0|acc |↑ |0.1931|± |0.0203|
124
+ | - high_school_biology | 1|none | 0|acc |↑ |0.2323|± |0.0240|
125
+ | - high_school_chemistry | 1|none | 0|acc |↑ |0.2266|± |0.0295|
126
+ | - high_school_computer_science | 1|none | 0|acc |↑ |0.2400|± |0.0429|
127
+ | - high_school_mathematics | 1|none | 0|acc |↑ |0.2037|± |0.0246|
128
+ | - high_school_physics | 1|none | 0|acc |↑ |0.2185|± |0.0337|
129
+ | - high_school_statistics | 1|none | 0|acc |↑ |0.1898|± |0.0267|
130
+ | - machine_learning | 1|none | 0|acc |↑ |0.3393|± |0.0449|
131
+ |truthfulqa_mc2 | 2|none | 0|acc |↑ |0.5061|± |0.0167|
132
+ |winogrande | 1|none | 0|acc |↑ |0.4933|± |0.0141|
133
+
134
+ | Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
135
+ |------------------|------:|------|------|------|---|-----:|---|-----:|
136
+ |mmlu | 2|none | |acc |↑ |0.2464|± |0.0036|
137
+ | - humanities | 2|none | |acc |↑ |0.2485|± |0.0063|
138
+ | - other | 2|none | |acc |↑ |0.2536|± |0.0078|
139
+ | - social sciences| 2|none | |acc |↑ |0.2535|± |0.0079|
140
+ | - stem | 2|none | |acc |↑ |0.2293|± |0.0075|
141
+
142
  ```bash
143
  litgpt evaluate --tasks 'leaderboard' --out_dir 'evaluate-leaderboard/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
144
  ```
145
 
146
+ | Tasks |Version|Filter|n-shot| Metric | |Value | |Stderr|
147
+ |-----------------------------------------------------------|-------|------|-----:|-----------------------|---|-----:|---|------|
148
+ |leaderboard | N/A| | | | | | | |
149
+ | - leaderboard_bbh | N/A| | | | | | | |
150
+ | - leaderboard_bbh_boolean_expressions | 1|none | 3|acc_norm |↑ |0.4600|± |0.0316|
151
+ | - leaderboard_bbh_causal_judgement | 1|none | 3|acc_norm |↑ |0.5134|± |0.0366|
152
+ | - leaderboard_bbh_date_understanding | 1|none | 3|acc_norm |↑ |0.1360|± |0.0217|
153
+ | - leaderboard_bbh_disambiguation_qa | 1|none | 3|acc_norm |↑ |0.2960|± |0.0289|
154
+ | - leaderboard_bbh_formal_fallacies | 1|none | 3|acc_norm |↑ |0.4760|± |0.0316|
155
+ | - leaderboard_bbh_geometric_shapes | 1|none | 3|acc_norm |↑ |0.0800|± |0.0172|
156
+ | - leaderboard_bbh_hyperbaton | 1|none | 3|acc_norm |↑ |0.5120|± |0.0317|
157
+ | - leaderboard_bbh_logical_deduction_five_objects | 1|none | 3|acc_norm |↑ |0.1760|± |0.0241|
158
+ | - leaderboard_bbh_logical_deduction_seven_objects | 1|none | 3|acc_norm |↑ |0.1320|± |0.0215|
159
+ | - leaderboard_bbh_logical_deduction_three_objects | 1|none | 3|acc_norm |↑ |0.3160|± |0.0295|
160
+ | - leaderboard_bbh_movie_recommendation | 1|none | 3|acc_norm |↑ |0.2480|± |0.0274|
161
+ | - leaderboard_bbh_navigate | 1|none | 3|acc_norm |↑ |0.4200|± |0.0313|
162
+ | - leaderboard_bbh_object_counting | 1|none | 3|acc_norm |↑ |0.0360|± |0.0118|
163
+ | - leaderboard_bbh_penguins_in_a_table | 1|none | 3|acc_norm |↑ |0.1986|± |0.0331|
164
+ | - leaderboard_bbh_reasoning_about_colored_objects | 1|none | 3|acc_norm |↑ |0.0520|± |0.0141|
165
+ | - leaderboard_bbh_ruin_names | 1|none | 3|acc_norm |↑ |0.2760|± |0.0283|
166
+ | - leaderboard_bbh_salient_translation_error_detection | 1|none | 3|acc_norm |↑ |0.1400|± |0.0220|
167
+ | - leaderboard_bbh_snarks | 1|none | 3|acc_norm |↑ |0.4326|± |0.0372|
168
+ | - leaderboard_bbh_sports_understanding | 1|none | 3|acc_norm |↑ |0.4600|± |0.0316|
169
+ | - leaderboard_bbh_temporal_sequences | 1|none | 3|acc_norm |↑ |0.2680|± |0.0281|
170
+ | - leaderboard_bbh_tracking_shuffled_objects_five_objects | 1|none | 3|acc_norm |↑ |0.2040|± |0.0255|
171
+ | - leaderboard_bbh_tracking_shuffled_objects_seven_objects| 1|none | 3|acc_norm |↑ |0.1640|± |0.0235|
172
+ | - leaderboard_bbh_tracking_shuffled_objects_three_objects| 1|none | 3|acc_norm |↑ |0.3840|± |0.0308|
173
+ | - leaderboard_bbh_web_of_lies | 1|none | 3|acc_norm |↑ |0.4880|± |0.0317|
174
+ | - leaderboard_gpqa | N/A| | | | | | | |
175
+ | - leaderboard_gpqa_diamond | 1|none | 0|acc_norm |↑ |0.2778|± |0.0319|
176
+ | - leaderboard_gpqa_extended | 1|none | 0|acc_norm |↑ |0.2766|± |0.0192|
177
+ | - leaderboard_gpqa_main | 1|none | 0|acc_norm |↑ |0.2031|± |0.0190|
178
+ | - leaderboard_ifeval | 3|none | 0|inst_level_loose_acc |↑ |0.1811|± | N/A|
179
+ | | |none | 0|inst_level_strict_acc |↑ |0.1715|± | N/A|
180
+ | | |none | 0|prompt_level_loose_acc |↑ |0.1091|± |0.0134|
181
+ | | |none | 0|prompt_level_strict_acc|↑ |0.1035|± |0.0131|
182
+ | - leaderboard_math_hard | N/A| | | | | | | |
183
+ | - leaderboard_math_algebra_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
184
+ | - leaderboard_math_counting_and_prob_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
185
+ | - leaderboard_math_geometry_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
186
+ | - leaderboard_math_intermediate_algebra_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
187
+ | - leaderboard_math_num_theory_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
188
+ | - leaderboard_math_prealgebra_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
189
+ | - leaderboard_math_precalculus_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
190
+ | - leaderboard_mmlu_pro | 0.1|none | 5|acc |↑ |0.1169|± |0.0029|
191
+ | - leaderboard_musr | N/A| | | | | | | |
192
+ | - leaderboard_musr_murder_mysteries | 1|none | 0|acc_norm |↑ |0.5080|± |0.0317|
193
+ | - leaderboard_musr_object_placements | 1|none | 0|acc_norm |↑ |0.3008|± |0.0287|
194
+ | - leaderboard_musr_team_allocation | 1|none | 0|acc_norm |↑ |0.3760|± |0.0307|
195
+
196
  ```bash
197
+ litgpt evaluate --tasks 'gsm8k,mathqa' --out_dir 'evaluate-math/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
198
  ```
199
 
200
+ |Tasks |Version| Filter |n-shot| Metric | |Value | |Stderr|
201
+ |------|------:|----------------|-----:|-----------|---|-----:|---|-----:|
202
+ |gsm8k | 3|flexible-extract| 5|exact_match|↑ |0.0190|± |0.0038|
203
+ | | |strict-match | 5|exact_match|↑ |0.0000|± |0.0000|
204
+ |mathqa| 1|none | 0|acc |↑ |0.2060|± |0.0074|
205
+ | | |none | 0|acc_norm |↑ |0.2057|± |0.0074|
206
+
207
  ```bash
208
  litgpt evaluate --tasks 'mmlu,mmlu_pro' --out_dir 'evaluate-mmlu/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
209
  ```
 
212
  litgpt evaluate --tasks 'arc_challenge,boolq,gpqa,hellaswag,openbookqa,piqa,truthfulqa_mc2,winogrande' --out_dir 'evaluate-reasoning/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
213
  ```
214
 
215
+ | Tasks |Version| Filter |n-shot| Metric | |Value | |Stderr|
216
+ |-------------------------------|------:|----------------|-----:|-----------|---|-----:|---|-----:|
217
+ |arc_challenge | 1|none | 0|acc |↑ |0.2176|± |0.0121|
218
+ | | |none | 0|acc_norm |↑ |0.2560|± |0.0128|
219
+ |boolq | 2|none | 0|acc |↑ |0.3783|± |0.0085|
220
+ |gpqa_diamond_cot_n_shot | 2|flexible-extract| 0|exact_match|↑ |0.0051|± |0.0051|
221
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
222
+ |gpqa_diamond_cot_zeroshot | 1|flexible-extract| 0|exact_match|↑ |0.0051|± |0.0051|
223
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
224
+ |gpqa_diamond_generative_n_shot | 2|flexible-extract| 0|exact_match|↑ |0.0051|± |0.0051|
225
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
226
+ |gpqa_diamond_n_shot | 2|none | 0|acc |↑ |0.1970|± |0.0283|
227
+ | | |none | 0|acc_norm |↑ |0.1970|± |0.0283|
228
+ |gpqa_diamond_zeroshot | 1|none | 0|acc |↑ |0.2727|± |0.0317|
229
+ | | |none | 0|acc_norm |↑ |0.2727|± |0.0317|
230
+ |gpqa_extended_cot_n_shot | 2|flexible-extract| 0|exact_match|↑ |0.0018|± |0.0018|
231
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
232
+ |gpqa_extended_cot_zeroshot | 1|flexible-extract| 0|exact_match|↑ |0.0037|± |0.0026|
233
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
234
+ |gpqa_extended_generative_n_shot| 2|flexible-extract| 0|exact_match|↑ |0.0073|± |0.0037|
235
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
236
+ |gpqa_extended_n_shot | 2|none | 0|acc |↑ |0.2564|± |0.0187|
237
+ | | |none | 0|acc_norm |↑ |0.2564|± |0.0187|
238
+ |gpqa_extended_zeroshot | 1|none | 0|acc |↑ |0.2802|± |0.0192|
239
+ | | |none | 0|acc_norm |↑ |0.2802|± |0.0192|
240
+ |gpqa_main_cot_n_shot | 2|flexible-extract| 0|exact_match|↑ |0.0000|± |0.0000|
241
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
242
+ |gpqa_main_cot_zeroshot | 1|flexible-extract| 0|exact_match|↑ |0.0000|± |0.0000|
243
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
244
+ |gpqa_main_generative_n_shot | 2|flexible-extract| 0|exact_match|↑ |0.0089|± |0.0044|
245
+ | | |strict-match | 0|exact_match|↑ |0.0000|± |0.0000|
246
+ |gpqa_main_n_shot | 2|none | 0|acc |↑ |0.2478|± |0.0204|
247
+ | | |none | 0|acc_norm |↑ |0.2478|± |0.0204|
248
+ |gpqa_main_zeroshot | 1|none | 0|acc |↑ |0.2143|± |0.0194|
249
+ | | |none | 0|acc_norm |↑ |0.2143|± |0.0194|
250
+ |hellaswag | 1|none | 0|acc |↑ |0.2618|± |0.0044|
251
+ | | |none | 0|acc_norm |↑ |0.2592|± |0.0044|
252
+ |openbookqa | 1|none | 0|acc |↑ |0.1340|± |0.0152|
253
+ | | |none | 0|acc_norm |↑ |0.2340|± |0.0190|
254
+ |piqa | 1|none | 0|acc |↑ |0.5201|± |0.0117|
255
+ | | |none | 0|acc_norm |↑ |0.5076|± |0.0117|
256
+ |truthfulqa_mc2 | 2|none | 0|acc |↑ |0.5061|± |0.0167|
257
+ |winogrande | 1|none | 0|acc |↑ |0.4933|± |0.0141|
258
 
259
  ```bash
260
+ litgpt evaluate --tasks 'mmlu_multilingual,mgsm' --out_dir 'evaluate-multilinguals/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
261
  ```
262
 
263
  ```bash
scripts/pretrain-model.yaml CHANGED
@@ -24,11 +24,11 @@ model_config:
24
  mlp_class_name: "LLaMAMLP"
25
  intermediate_size: 2048
26
  rope_base: 500000
27
- # rope_adjustments:
28
- # factor: 32.0
29
- # low_freq_factor: 1.0
30
- # high_freq_factor: 4.0
31
- # original_max_seq_len: 8192
32
 
33
  # Directory in which to save checkpoints and logs. If running in a Lightning Studio Job, look for it in
34
  # /teamspace/jobs/<job-name>/share. (type: <class 'Path'>, default: out/pretrain)
@@ -120,14 +120,14 @@ optimizer:
120
  class_path: grokadamw.GrokAdamW
121
  # class_path: bitsandbytes.optim.AdamW8bit
122
  # class_path: bitsandbytes.optim.PagedAdamW8bit
123
-
124
  init_args:
125
  # (type: float, default: 0.001)
126
  lr: 1e-3
127
-
128
  # (type: float, default: 0.01)
129
  weight_decay: 1e-2
130
-
131
  # (type: tuple, default: (0.9,0.999))
132
  betas:
133
  - 0.9
 
24
  mlp_class_name: "LLaMAMLP"
25
  intermediate_size: 2048
26
  rope_base: 500000
27
+ rope_adjustments:
28
+ factor: 32.0
29
+ low_freq_factor: 1.0
30
+ high_freq_factor: 4.0
31
+ original_max_seq_len: 8192
32
 
33
  # Directory in which to save checkpoints and logs. If running in a Lightning Studio Job, look for it in
34
  # /teamspace/jobs/<job-name>/share. (type: <class 'Path'>, default: out/pretrain)
 
120
  class_path: grokadamw.GrokAdamW
121
  # class_path: bitsandbytes.optim.AdamW8bit
122
  # class_path: bitsandbytes.optim.PagedAdamW8bit
123
+
124
  init_args:
125
  # (type: float, default: 0.001)
126
  lr: 1e-3
127
+
128
  # (type: float, default: 0.01)
129
  weight_decay: 1e-2
130
+
131
  # (type: tuple, default: (0.9,0.999))
132
  betas:
133
  - 0.9