cleanup
Browse files- README.md +146 -1
- scripts/TRAIN.md +4 -2
README.md
CHANGED
|
@@ -26,6 +26,12 @@ tags:
|
|
| 26 |
|
| 27 |
|
| 28 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 29 |
[loss, val_loss](https://api.wandb.ai/links/mtasic85/ecf4l9qp)
|
| 30 |
|
| 31 |
[val_ppl](https://api.wandb.ai/links/mtasic85/qsn8mz13)
|
|
@@ -34,4 +40,143 @@ tags:
|
|
| 34 |
|
| 35 |
[learning_rate](https://api.wandb.ai/links/mtasic85/7kyopu4t)
|
| 36 |
|
| 37 |
-
## lm-evaluation-harness
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
|
| 28 |
|
| 29 |
+
A pretrained language model based on the Llama model with about **109M** parameters. This model has been trained on **9.7B** (`9,782,206,713`) tokens from more than **5.2M** (`5,285,575`) dataset rows.
|
| 30 |
+
|
| 31 |
+
This model **isn't** designed for immediate use but rather for Continued Pretraining and Finetuning on a downstream task. While it can handle a context length of up to **32K** (`32,768`) tokens, it was pretrained with sequences of **2K** (`2048`) tokens.
|
| 32 |
+
|
| 33 |
+
The objective is to streamline the cognitive or reasoning core, eliminating any redundant knowledge from the model.
|
| 34 |
+
|
| 35 |
[loss, val_loss](https://api.wandb.ai/links/mtasic85/ecf4l9qp)
|
| 36 |
|
| 37 |
[val_ppl](https://api.wandb.ai/links/mtasic85/qsn8mz13)
|
|
|
|
| 40 |
|
| 41 |
[learning_rate](https://api.wandb.ai/links/mtasic85/7kyopu4t)
|
| 42 |
|
| 43 |
+
## lm-evaluation-harness
|
| 44 |
+
|
| 45 |
+
```bash
|
| 46 |
+
litgpt evaluate --tasks 'leaderboard' --out_dir 'evaluate-0/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
|
| 47 |
+
```
|
| 48 |
+
|
| 49 |
+
| Tasks |Version|Filter|n-shot| Metric | |Value | |Stderr|
|
| 50 |
+
|-----------------------------------------------------------|-------|------|-----:|-----------------------|---|-----:|---|------|
|
| 51 |
+
|leaderboard | N/A| | | | | | | |
|
| 52 |
+
| - leaderboard_bbh | N/A| | | | | | | |
|
| 53 |
+
| - leaderboard_bbh_boolean_expressions | 1|none | 3|acc_norm |↑ |0.5680|± |0.0314|
|
| 54 |
+
| - leaderboard_bbh_causal_judgement | 1|none | 3|acc_norm |↑ |0.5294|± |0.0366|
|
| 55 |
+
| - leaderboard_bbh_date_understanding | 1|none | 3|acc_norm |↑ |0.1880|± |0.0248|
|
| 56 |
+
| - leaderboard_bbh_disambiguation_qa | 1|none | 3|acc_norm |↑ |0.3240|± |0.0297|
|
| 57 |
+
| - leaderboard_bbh_formal_fallacies | 1|none | 3|acc_norm |↑ |0.4720|± |0.0316|
|
| 58 |
+
| - leaderboard_bbh_geometric_shapes | 1|none | 3|acc_norm |↑ |0.0280|± |0.0105|
|
| 59 |
+
| - leaderboard_bbh_hyperbaton | 1|none | 3|acc_norm |↑ |0.5160|± |0.0317|
|
| 60 |
+
| - leaderboard_bbh_logical_deduction_five_objects | 1|none | 3|acc_norm |↑ |0.1760|± |0.0241|
|
| 61 |
+
| - leaderboard_bbh_logical_deduction_seven_objects | 1|none | 3|acc_norm |↑ |0.1360|± |0.0217|
|
| 62 |
+
| - leaderboard_bbh_logical_deduction_three_objects | 1|none | 3|acc_norm |↑ |0.3480|± |0.0302|
|
| 63 |
+
| - leaderboard_bbh_movie_recommendation | 1|none | 3|acc_norm |↑ |0.2280|± |0.0266|
|
| 64 |
+
| - leaderboard_bbh_navigate | 1|none | 3|acc_norm |↑ |0.4800|± |0.0317|
|
| 65 |
+
| - leaderboard_bbh_object_counting | 1|none | 3|acc_norm |↑ |0.0640|± |0.0155|
|
| 66 |
+
| - leaderboard_bbh_penguins_in_a_table | 1|none | 3|acc_norm |↑ |0.2329|± |0.0351|
|
| 67 |
+
| - leaderboard_bbh_reasoning_about_colored_objects | 1|none | 3|acc_norm |↑ |0.1240|± |0.0209|
|
| 68 |
+
| - leaderboard_bbh_ruin_names | 1|none | 3|acc_norm |↑ |0.2400|± |0.0271|
|
| 69 |
+
| - leaderboard_bbh_salient_translation_error_detection | 1|none | 3|acc_norm |↑ |0.1560|± |0.0230|
|
| 70 |
+
| - leaderboard_bbh_snarks | 1|none | 3|acc_norm |↑ |0.4607|± |0.0375|
|
| 71 |
+
| - leaderboard_bbh_sports_understanding | 1|none | 3|acc_norm |↑ |0.4560|± |0.0316|
|
| 72 |
+
| - leaderboard_bbh_temporal_sequences | 1|none | 3|acc_norm |↑ |0.2320|± |0.0268|
|
| 73 |
+
| - leaderboard_bbh_tracking_shuffled_objects_five_objects | 1|none | 3|acc_norm |↑ |0.2000|± |0.0253|
|
| 74 |
+
| - leaderboard_bbh_tracking_shuffled_objects_seven_objects| 1|none | 3|acc_norm |↑ |0.1520|± |0.0228|
|
| 75 |
+
| - leaderboard_bbh_tracking_shuffled_objects_three_objects| 1|none | 3|acc_norm |↑ |0.3160|± |0.0295|
|
| 76 |
+
| - leaderboard_bbh_web_of_lies | 1|none | 3|acc_norm |↑ |0.5040|± |0.0317|
|
| 77 |
+
| - leaderboard_gpqa | N/A| | | | | | | |
|
| 78 |
+
| - leaderboard_gpqa_diamond | 1|none | 0|acc_norm |↑ |0.1919|± |0.0281|
|
| 79 |
+
| - leaderboard_gpqa_extended | 1|none | 0|acc_norm |↑ |0.2747|± |0.0191|
|
| 80 |
+
| - leaderboard_gpqa_main | 1|none | 0|acc_norm |↑ |0.2589|± |0.0207|
|
| 81 |
+
| - leaderboard_ifeval | 3|none | 0|inst_level_loose_acc |↑ |0.2002|± | N/A|
|
| 82 |
+
| | |none | 0|inst_level_strict_acc |↑ |0.1871|± | N/A|
|
| 83 |
+
| | |none | 0|prompt_level_loose_acc |↑ |0.1072|± |0.0133|
|
| 84 |
+
| | |none | 0|prompt_level_strict_acc|↑ |0.0998|± |0.0129|
|
| 85 |
+
| - leaderboard_math_hard | N/A| | | | | | | |
|
| 86 |
+
| - leaderboard_math_algebra_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
|
| 87 |
+
| - leaderboard_math_counting_and_prob_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
|
| 88 |
+
| - leaderboard_math_geometry_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
|
| 89 |
+
| - leaderboard_math_intermediate_algebra_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
|
| 90 |
+
| - leaderboard_math_num_theory_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
|
| 91 |
+
| - leaderboard_math_prealgebra_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
|
| 92 |
+
| - leaderboard_math_precalculus_hard | 1|none | 4|exact_match |↑ |0.0000|± | 0|
|
| 93 |
+
| - leaderboard_mmlu_pro | 0.1|none | 5|acc |↑ |0.1096|± |0.0028|
|
| 94 |
+
| - leaderboard_musr | N/A| | | | | | | |
|
| 95 |
+
| - leaderboard_musr_murder_mysteries | 1|none | 0|acc_norm |↑ |0.4800|± |0.0317|
|
| 96 |
+
| - leaderboard_musr_object_placements | 1|none | 0|acc_norm |↑ |0.2930|± |0.0285|
|
| 97 |
+
| - leaderboard_musr_team_allocation | 1|none | 0|acc_norm |↑ |0.3360|± |0.0299|
|
| 98 |
+
|
| 99 |
+
```bash
|
| 100 |
+
litgpt evaluate --tasks 'hellaswag,gsm8k,truthfulqa_mc2,mmlu,winogrande,arc_challenge' --out_dir 'evaluate-1/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
|
| 101 |
+
```
|
| 102 |
+
|
| 103 |
+
Tasks |Version| Filter |n-shot| Metric | |Value | |Stderr|
|
| 104 |
+
|---------------------------------------|------:|----------------|-----:|-----------|---|-----:|---|-----:|
|
| 105 |
+
|arc_challenge | 1|none | 0|acc |↑ |0.2082|± |0.0119|
|
| 106 |
+
| | |none | 0|acc_norm |↑ |0.2474|± |0.0126|
|
| 107 |
+
|gsm8k | 3|flexible-extract| 5|exact_match|↑ |0.0106|± |0.0028|
|
| 108 |
+
| | |strict-match | 5|exact_match|↑ |0.0008|± |0.0008|
|
| 109 |
+
|hellaswag | 1|none | 0|acc |↑ |0.2766|± |0.0045|
|
| 110 |
+
| | |none | 0|acc_norm |↑ |0.2926|± |0.0045|
|
| 111 |
+
|mmlu | 2|none | |acc |↑ |0.2349|± |0.0036|
|
| 112 |
+
| - humanities | 2|none | |acc |↑ |0.2461|± |0.0063|
|
| 113 |
+
| - formal_logic | 1|none | 0|acc |↑ |0.2698|± |0.0397|
|
| 114 |
+
| - high_school_european_history | 1|none | 0|acc |↑ |0.2000|± |0.0312|
|
| 115 |
+
| - high_school_us_history | 1|none | 0|acc |↑ |0.2549|± |0.0306|
|
| 116 |
+
| - high_school_world_history | 1|none | 0|acc |↑ |0.2616|± |0.0286|
|
| 117 |
+
| - international_law | 1|none | 0|acc |↑ |0.2479|± |0.0394|
|
| 118 |
+
| - jurisprudence | 1|none | 0|acc |↑ |0.2593|± |0.0424|
|
| 119 |
+
| - logical_fallacies | 1|none | 0|acc |↑ |0.2638|± |0.0346|
|
| 120 |
+
| - moral_disputes | 1|none | 0|acc |↑ |0.2457|± |0.0232|
|
| 121 |
+
| - moral_scenarios | 1|none | 0|acc |↑ |0.2458|± |0.0144|
|
| 122 |
+
| - philosophy | 1|none | 0|acc |↑ |0.1833|± |0.0220|
|
| 123 |
+
| - prehistory | 1|none | 0|acc |↑ |0.2315|± |0.0235|
|
| 124 |
+
| - professional_law | 1|none | 0|acc |↑ |0.2503|± |0.0111|
|
| 125 |
+
| - world_religions | 1|none | 0|acc |↑ |0.3216|± |0.0358|
|
| 126 |
+
| - other | 2|none | |acc |↑ |0.2391|± |0.0076|
|
| 127 |
+
| - business_ethics | 1|none | 0|acc |↑ |0.2900|± |0.0456|
|
| 128 |
+
| - clinical_knowledge | 1|none | 0|acc |↑ |0.2377|± |0.0262|
|
| 129 |
+
| - college_medicine | 1|none | 0|acc |↑ |0.2197|± |0.0316|
|
| 130 |
+
| - global_facts | 1|none | 0|acc |↑ |0.2100|± |0.0409|
|
| 131 |
+
| - human_aging | 1|none | 0|acc |↑ |0.2960|± |0.0306|
|
| 132 |
+
| - management | 1|none | 0|acc |↑ |0.1748|± |0.0376|
|
| 133 |
+
| - marketing | 1|none | 0|acc |↑ |0.2949|± |0.0299|
|
| 134 |
+
| - medical_genetics | 1|none | 0|acc |↑ |0.2700|± |0.0446|
|
| 135 |
+
| - miscellaneous | 1|none | 0|acc |↑ |0.2222|± |0.0149|
|
| 136 |
+
| - nutrition | 1|none | 0|acc |↑ |0.2092|± |0.0233|
|
| 137 |
+
| - professional_accounting | 1|none | 0|acc |↑ |0.2518|± |0.0259|
|
| 138 |
+
| - professional_medicine | 1|none | 0|acc |↑ |0.1949|± |0.0241|
|
| 139 |
+
| - virology | 1|none | 0|acc |↑ |0.3012|± |0.0357|
|
| 140 |
+
| - social sciences | 2|none | |acc |↑ |0.2246|± |0.0075|
|
| 141 |
+
| - econometrics | 1|none | 0|acc |↑ |0.2807|± |0.0423|
|
| 142 |
+
| - high_school_geography | 1|none | 0|acc |↑ |0.1818|± |0.0275|
|
| 143 |
+
| - high_school_government_and_politics| 1|none | 0|acc |↑ |0.2176|± |0.0298|
|
| 144 |
+
| - high_school_macroeconomics | 1|none | 0|acc |↑ |0.2179|± |0.0209|
|
| 145 |
+
| - high_school_microeconomics | 1|none | 0|acc |↑ |0.2101|± |0.0265|
|
| 146 |
+
| - high_school_psychology | 1|none | 0|acc |↑ |0.2000|± |0.0171|
|
| 147 |
+
| - human_sexuality | 1|none | 0|acc |↑ |0.2519|± |0.0381|
|
| 148 |
+
| - professional_psychology | 1|none | 0|acc |↑ |0.2516|± |0.0176|
|
| 149 |
+
| - public_relations | 1|none | 0|acc |↑ |0.2182|± |0.0396|
|
| 150 |
+
| - security_studies | 1|none | 0|acc |↑ |0.1959|± |0.0254|
|
| 151 |
+
| - sociology | 1|none | 0|acc |↑ |0.2488|± |0.0306|
|
| 152 |
+
| - us_foreign_policy | 1|none | 0|acc |↑ |0.2800|± |0.0451|
|
| 153 |
+
| - stem | 2|none | |acc |↑ |0.2239|± |0.0074|
|
| 154 |
+
| - abstract_algebra | 1|none | 0|acc |↑ |0.1800|± |0.0386|
|
| 155 |
+
| - anatomy | 1|none | 0|acc |↑ |0.1778|± |0.0330|
|
| 156 |
+
| - astronomy | 1|none | 0|acc |↑ |0.1974|± |0.0324|
|
| 157 |
+
| - college_biology | 1|none | 0|acc |↑ |0.2569|± |0.0365|
|
| 158 |
+
| - college_chemistry | 1|none | 0|acc |↑ |0.2400|± |0.0429|
|
| 159 |
+
| - college_computer_science | 1|none | 0|acc |↑ |0.2400|± |0.0429|
|
| 160 |
+
| - college_mathematics | 1|none | 0|acc |↑ |0.2400|± |0.0429|
|
| 161 |
+
| - college_physics | 1|none | 0|acc |↑ |0.2255|± |0.0416|
|
| 162 |
+
| - computer_security | 1|none | 0|acc |↑ |0.2700|± |0.0446|
|
| 163 |
+
| - conceptual_physics | 1|none | 0|acc |↑ |0.2468|± |0.0282|
|
| 164 |
+
| - electrical_engineering | 1|none | 0|acc |↑ |0.2552|± |0.0363|
|
| 165 |
+
| - elementary_mathematics | 1|none | 0|acc |↑ |0.2407|± |0.0220|
|
| 166 |
+
| - high_school_biology | 1|none | 0|acc |↑ |0.1710|± |0.0214|
|
| 167 |
+
| - high_school_chemistry | 1|none | 0|acc |↑ |0.1724|± |0.0266|
|
| 168 |
+
| - high_school_computer_science | 1|none | 0|acc |↑ |0.2600|± |0.0441|
|
| 169 |
+
| - high_school_mathematics | 1|none | 0|acc |↑ |0.2519|± |0.0265|
|
| 170 |
+
| - high_school_physics | 1|none | 0|acc |↑ |0.1457|± |0.0288|
|
| 171 |
+
| - high_school_statistics | 1|none | 0|acc |↑ |0.2083|± |0.0277|
|
| 172 |
+
| - machine_learning | 1|none | 0|acc |↑ |0.3571|± |0.0455|
|
| 173 |
+
|truthfulqa_mc2 | 2|none | 0|acc |↑ |0.4506|± |0.0161|
|
| 174 |
+
|winogrande | 1|none | 0|acc |↑ |0.5288|± |0.0140|
|
| 175 |
+
|
| 176 |
+
| Groups |Version|Filter|n-shot|Metric| |Value | |Stderr|
|
| 177 |
+
|------------------|------:|------|------|------|---|-----:|---|-----:|
|
| 178 |
+
|mmlu | 2|none | |acc |↑ |0.2349|± |0.0036|
|
| 179 |
+
| - humanities | 2|none | |acc |↑ |0.2461|± |0.0063|
|
| 180 |
+
| - other | 2|none | |acc |↑ |0.2391|± |0.0076|
|
| 181 |
+
| - social sciences| 2|none | |acc |↑ |0.2246|± |0.0075|
|
| 182 |
+
| - stem | 2|none | |acc |↑ |0.2239|± |0.0074|
|
scripts/TRAIN.md
CHANGED
|
@@ -57,7 +57,9 @@ model.save_pretrained('out/converted_model/')
|
|
| 57 |
## Evaluate
|
| 58 |
|
| 59 |
```bash
|
| 60 |
-
|
| 61 |
|
| 62 |
-
litgpt evaluate --tasks 'hellaswag,gsm8k,truthfulqa_mc2,mmlu,
|
|
|
|
|
|
|
| 63 |
```
|
|
|
|
| 57 |
## Evaluate
|
| 58 |
|
| 59 |
```bash
|
| 60 |
+
litgpt evaluate --tasks 'leaderboard' --out_dir 'evaluate-0/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
|
| 61 |
|
| 62 |
+
litgpt evaluate --tasks 'hellaswag,gsm8k,truthfulqa_mc2,mmlu,winogrande,arc_challenge' --out_dir 'evaluate-1/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
|
| 63 |
+
|
| 64 |
+
litgpt evaluate --tasks 'mmlu_pro,ifeval,mgsm_direct,mathqa,gpqa' --out_dir 'evaluate-2/' --batch_size 4 --dtype 'bfloat16' out/pretrain/final/
|
| 65 |
```
|