Add Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy step 114274 weights
Browse files- .gitattributes +33 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/README.md +9 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/code.diff +0 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/complete.json +21 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/config.json +112 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/config.yaml +48 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/.metadata +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__0_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__10_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__11_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__12_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__13_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__14_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__15_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__16_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__17_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__18_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__19_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__1_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__20_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__21_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__22_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__23_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__24_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__25_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__26_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__27_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__28_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__29_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__2_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__30_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__31_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__3_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__4_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__5_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__6_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__7_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__8_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__9_0.distcp +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/checkpoint_inventory.json +26 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/receipt.json +74 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/manifest.json +191 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/model.safetensors +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/original_model.safetensors +3 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/recipe.json +278 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/snapshot_manifest.json +497 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/tokenizer/tokenizer.json +0 -0
- small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/training_code.tar.gz +3 -0
.gitattributes
CHANGED
|
@@ -100,3 +100,36 @@ Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA-progressiveK-75
|
|
| 100 |
Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA-progressiveK-75-90--vyh309f5/full_epoch_3/checkpoints/fsdp2_epoch_3/__8_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 101 |
Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA-progressiveK-75-90--vyh309f5/full_epoch_3/checkpoints/fsdp2_epoch_3/__9_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 102 |
small_search_1epoch/evaluation/20261005-200230_small-checkpoints-h20-handoff-t4/HRM-Text.bundle filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 100 |
Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA-progressiveK-75-90--vyh309f5/full_epoch_3/checkpoints/fsdp2_epoch_3/__8_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 101 |
Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA-progressiveK-75-90--vyh309f5/full_epoch_3/checkpoints/fsdp2_epoch_3/__9_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 102 |
small_search_1epoch/evaluation/20261005-200230_small-checkpoints-h20-handoff-t4/HRM-Text.bundle filter=lfs diff=lfs merge=lfs -text
|
| 103 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/.metadata filter=lfs diff=lfs merge=lfs -text
|
| 104 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__0_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 105 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__10_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 106 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__11_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 107 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__12_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 108 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__13_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 109 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__14_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 110 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__15_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 111 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__16_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 112 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__17_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 113 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__18_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 114 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__19_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 115 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__1_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 116 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__20_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 117 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__21_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 118 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__22_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 119 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__23_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 120 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__24_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 121 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__25_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 122 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__26_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 123 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__27_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 124 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__28_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 125 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__29_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 126 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__2_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 127 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__30_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 128 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__31_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 129 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__3_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 130 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__4_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 131 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__5_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 132 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__6_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 133 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__7_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 134 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__8_0.distcp filter=lfs diff=lfs merge=lfs -text
|
| 135 |
+
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__9_0.distcp filter=lfs diff=lfs merge=lfs -text
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/README.md
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32 -- r4jyjwuy, epoch 1, step 114274
|
| 2 |
+
|
| 3 |
+
Final native checkpoint. model.safetensors contains EMA; original_model.safetensors contains raw model weights. Both preserve source keys/dtypes/values and passed finite/exact serialization checks.
|
| 4 |
+
|
| 5 |
+
full_checkpoint/ preserves all 32 original DCP shards, metadata and receipt, including optimizer moments, step counters and EMA. Do not load only rank 0. No standalone RNG/scheduler state was saved by training.
|
| 6 |
+
|
| 7 |
+
Use bundled matching training_code.tar.gz and saved recipe/config. K3 refined-pseudo feedback is not a Transformers AutoModel or K1 export. The official native run-root loader accepts the full checkpoint; build_lookahead_model accepts these explicit model/lookahead configurations and safetensors state. The legacy safetensors CLI loader's K1 guard must not be bypassed by changing K3 metadata.
|
| 8 |
+
|
| 9 |
+
Evaluation uses EMA with supported immutable fused feedback, two prompt traversals and one traversal per decoded token; trained maximum K=3. Original and extended prompt protocols remain separate.
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/code.diff
ADDED
|
File without changes
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/complete.json
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"status": "complete",
|
| 3 |
+
"partial_epoch": false,
|
| 4 |
+
"stop_at_unix_time": null,
|
| 5 |
+
"final_checkpoint": "/apdcephfs_tj31/share_303214315/hunyuan/kenanchen/loopTF-workspace/output/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32/checkpoints/fsdp2_epoch_1",
|
| 6 |
+
"epochs": 1,
|
| 7 |
+
"step": 114274,
|
| 8 |
+
"input_tokens": 14922536242,
|
| 9 |
+
"response_labels": 6402595422,
|
| 10 |
+
"examples": 57699115,
|
| 11 |
+
"elapsed_seconds": 102927.03982994054,
|
| 12 |
+
"pass_schedule": null,
|
| 13 |
+
"pass_accounting": {
|
| 14 |
+
"3": {
|
| 15 |
+
"updates": 114274,
|
| 16 |
+
"input_tokens": 14922536242,
|
| 17 |
+
"response_labels": 6402595422,
|
| 18 |
+
"examples": 57699115
|
| 19 |
+
}
|
| 20 |
+
}
|
| 21 |
+
}
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/config.json
ADDED
|
@@ -0,0 +1,112 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"checkpoint_format": "safetensors_native_hrm_lookahead_state_dict",
|
| 3 |
+
"architecture": "hrm_text_lookahead",
|
| 4 |
+
"weights": "ema",
|
| 5 |
+
"epoch": 1,
|
| 6 |
+
"iteration": 114274,
|
| 7 |
+
"parameter_count": 246672440,
|
| 8 |
+
"model_config": {
|
| 9 |
+
"model_type": "hrm_text",
|
| 10 |
+
"backbone_profile": "hrm_text",
|
| 11 |
+
"vocab_size": 65536,
|
| 12 |
+
"hidden_size": 760,
|
| 13 |
+
"intermediate_size": 2048,
|
| 14 |
+
"num_hidden_layers": 16,
|
| 15 |
+
"num_attention_heads": 10,
|
| 16 |
+
"num_key_value_heads": 10,
|
| 17 |
+
"max_position_embeddings": 4096,
|
| 18 |
+
"rope_theta": 10000.0,
|
| 19 |
+
"rms_norm_eps": 1e-06,
|
| 20 |
+
"initializer_range": 0.03627381250550058,
|
| 21 |
+
"attention_dropout": 0.0,
|
| 22 |
+
"tie_word_embeddings": false,
|
| 23 |
+
"gradient_checkpointing": false,
|
| 24 |
+
"padded_vocab_size": null,
|
| 25 |
+
"paper_sliding_window": 0,
|
| 26 |
+
"paper_full_attention_interval": 6,
|
| 27 |
+
"paper_allow_ffn_override": false,
|
| 28 |
+
"pretrained_model_name_or_path": null,
|
| 29 |
+
"revision": null,
|
| 30 |
+
"trust_remote_code": false,
|
| 31 |
+
"attn_implementation": "sdpa",
|
| 32 |
+
"hf_config_overrides": {},
|
| 33 |
+
"backbone_layout": "hrm_cycle",
|
| 34 |
+
"pseudo_low_source": "real_final",
|
| 35 |
+
"l_layers": 8,
|
| 36 |
+
"l_cycles": 3,
|
| 37 |
+
"h_layers": 8,
|
| 38 |
+
"h_cycles": 2,
|
| 39 |
+
"refiner_repeats": 2
|
| 40 |
+
},
|
| 41 |
+
"lookahead_config": {
|
| 42 |
+
"mode": "parallel",
|
| 43 |
+
"num_passes": 3,
|
| 44 |
+
"feedback_loss_weight": 1.0,
|
| 45 |
+
"attention_backend": "triton",
|
| 46 |
+
"checkpoint_refiner": false,
|
| 47 |
+
"cache_layer_indices": null,
|
| 48 |
+
"num_refiner_layers": 3,
|
| 49 |
+
"refiner_activation": "swiglu",
|
| 50 |
+
"refiner_intermediate_size": null,
|
| 51 |
+
"refiner_gate_bias": -1.0,
|
| 52 |
+
"future_loss_weight": 0.2,
|
| 53 |
+
"future_gradient_mode": "all",
|
| 54 |
+
"draft_loss_weight": 0,
|
| 55 |
+
"auxiliary_loss_passes": "all",
|
| 56 |
+
"auxiliary_loss_weighting": "fbw",
|
| 57 |
+
"refined_future_loss_weight": 0.0,
|
| 58 |
+
"pseudo_initializer": "A2",
|
| 59 |
+
"pseudo_candidate_mode": "soft_embedding",
|
| 60 |
+
"pseudo_top_k": 128,
|
| 61 |
+
"pseudo_temperature": 1.0,
|
| 62 |
+
"normalize_feedback_embedding": false,
|
| 63 |
+
"normalize_feedback_output": false,
|
| 64 |
+
"feedback_source": "refined_pseudo",
|
| 65 |
+
"refiner_variant": "R0",
|
| 66 |
+
"latent_horizon": 2,
|
| 67 |
+
"thinking_positions": 1,
|
| 68 |
+
"thinking_selection": "first_response",
|
| 69 |
+
"refiner_memory": true,
|
| 70 |
+
"refiner_memory_only": true,
|
| 71 |
+
"allow_objective_change_on_resume": false
|
| 72 |
+
},
|
| 73 |
+
"trained_active_passes": 3,
|
| 74 |
+
"upstream_revision": "aaa948ea674fd84b7bc455c9cfb455ecfefdf914",
|
| 75 |
+
"tokenizer_info": {
|
| 76 |
+
"tokenizer_path": "tokenizer/tokenizer.json",
|
| 77 |
+
"boq": "<|im_start|>",
|
| 78 |
+
"eoq": "<|im_end|>",
|
| 79 |
+
"eoa": "<|box_end|>",
|
| 80 |
+
"condition_mapping": {
|
| 81 |
+
"cot": "<|object_ref_end|>",
|
| 82 |
+
"synth": "<|quad_end|>",
|
| 83 |
+
"direct": "<|object_ref_start|>",
|
| 84 |
+
"noisy": "<|quad_start|>"
|
| 85 |
+
},
|
| 86 |
+
"vocab_size": 65536
|
| 87 |
+
},
|
| 88 |
+
"run_name": "Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32",
|
| 89 |
+
"wandb_id": "r4jyjwuy",
|
| 90 |
+
"training": {
|
| 91 |
+
"status": "complete",
|
| 92 |
+
"partial_epoch": false,
|
| 93 |
+
"stop_at_unix_time": null,
|
| 94 |
+
"final_checkpoint": "/apdcephfs_tj31/share_303214315/hunyuan/kenanchen/loopTF-workspace/output/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32/checkpoints/fsdp2_epoch_1",
|
| 95 |
+
"epochs": 1,
|
| 96 |
+
"step": 114274,
|
| 97 |
+
"input_tokens": 14922536242,
|
| 98 |
+
"response_labels": 6402595422,
|
| 99 |
+
"examples": 57699115,
|
| 100 |
+
"elapsed_seconds": 102927.03982994054,
|
| 101 |
+
"pass_schedule": null,
|
| 102 |
+
"pass_accounting": {
|
| 103 |
+
"3": {
|
| 104 |
+
"updates": 114274,
|
| 105 |
+
"input_tokens": 14922536242,
|
| 106 |
+
"response_labels": 6402595422,
|
| 107 |
+
"examples": 57699115
|
| 108 |
+
}
|
| 109 |
+
}
|
| 110 |
+
},
|
| 111 |
+
"training_code_revision": "7cf1e58e0f1b7049bf243da93bb9fdc961864449"
|
| 112 |
+
}
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/config.yaml
ADDED
|
@@ -0,0 +1,48 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
arch:
|
| 2 |
+
H_cycles: 2
|
| 3 |
+
H_override:
|
| 4 |
+
n_layers: 8
|
| 5 |
+
L_cycles: 3
|
| 6 |
+
attn_type: causal
|
| 7 |
+
expansion: 4
|
| 8 |
+
half_layers: false
|
| 9 |
+
hidden_size: 760
|
| 10 |
+
n_layers: 8
|
| 11 |
+
num_heads: 10
|
| 12 |
+
checkpoint_path: /root/dl/runtime/script/loopTF-workspace/output/causal_hrm_text/small_search/double_compute/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch
|
| 13 |
+
data:
|
| 14 |
+
path: /root/dl/runtime/script/loopTF-workspace/data/causal_hrm_text_sources/sampled
|
| 15 |
+
defaults:
|
| 16 |
+
- cfg_pretrain
|
| 17 |
+
- _self_
|
| 18 |
+
epochs: 1
|
| 19 |
+
experiment:
|
| 20 |
+
arm: lookahead
|
| 21 |
+
lookahead:
|
| 22 |
+
backbone_layout: hrm_cycle
|
| 23 |
+
feedback_source: refined_pseudo
|
| 24 |
+
h_cycles: 2
|
| 25 |
+
num_passes: 3
|
| 26 |
+
num_refiner_layers: 3
|
| 27 |
+
pseudo_low_source: real_final
|
| 28 |
+
refiner_gate_bias: -1.0
|
| 29 |
+
refiner_repeats: 2
|
| 30 |
+
losses:
|
| 31 |
+
auxiliary_loss_passes: all
|
| 32 |
+
auxiliary_loss_weighting: fbw
|
| 33 |
+
draft_loss_weight: 0
|
| 34 |
+
future_loss_weight: 0.2
|
| 35 |
+
global_batch_size: 131072
|
| 36 |
+
hydra:
|
| 37 |
+
run:
|
| 38 |
+
dir: /root/dl/runtime/script/loopTF-workspace/output/causal_hrm_text/small_search/double_compute/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch
|
| 39 |
+
lr: 0.0003
|
| 40 |
+
lr_warmup_steps: 750
|
| 41 |
+
project_name: lookahead_hrm_small_search
|
| 42 |
+
recipe_profile: small_search_1epoch
|
| 43 |
+
reference_validation:
|
| 44 |
+
capacity: 4096
|
| 45 |
+
every: 1000
|
| 46 |
+
suite: /root/dl/runtime/script/loopTF-workspace/data/evaluation/hrm_text/reference-v1-seed20261004
|
| 47 |
+
weights: ema
|
| 48 |
+
run_name: Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/.metadata
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b97f4d041169fa503d9cdf05793d8577f82d990add2c4a9d2c2a118bd0bc9451
|
| 3 |
+
size 3914233
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__0_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:98c78e37ea6c2c9927517422b826d672b78f821b5a9bfc998591245b5e50e424
|
| 3 |
+
size 124811113
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__10_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7fbb54c0bfdec89cec63946da7cce5c710c727cdf187319e096303046ff26b17
|
| 3 |
+
size 124824058
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__11_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:386ea6f23ae99b33689d6ce799a749d5cd9247857889a862c3365c1079e2fc12
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__12_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:577f7c603621263cd49e94e76bbaf23f10a14b56643d1e6a2f109fc251b8ca04
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__13_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a19f429adb42c7a147c086d697c1fe489590f721ef706b2d95abcb33e2fccd10
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__14_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:13ebd9be44bb65dc2bc7ef3c6fcae44cf53b59efe101b49ef4756651842ce032
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__15_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1028ad750e6eb5cd1c4ce4de9e85897ab16da4199c8852c2ca9ff4ffcbecd2b4
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__16_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a037c8e382b239d3a1a87666ca4892cdaffecb9b35a5dddd2899c4373d9523db
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__17_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ea6d72044c2fb61d26812001fd78803a4e707079e217ef2a63119962261678a
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__18_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8ff03bbc6d0a974dfc75b2b14c3a305e8851536172bb158fd63c794c73a88637
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__19_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:031e6bc9958fa694f9249d86497b398331044bb114eff9abeee7afb1ec31bd06
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__1_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:daf585fcfa65f3927e8c02a0c7ab5f8e46a9ba280d61e3b959b6bdd130973a49
|
| 3 |
+
size 124817526
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__20_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:56e7c51587ff1bde1cfbc3e2492dc28216199ef9f57a0ac0b2c66f6fdd875d85
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__21_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e52b270e8abb3858fd676c352120772401e8298a53db680c438a548f302e74ab
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__22_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e44acb5b968124b89fc6011448a7bc5e8b2b92a6cf71c28d26b373ded032cd5b
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__23_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f1474e6bf7cd4209a951e44d93ee0014eef4ea988b476612438fed67d8270b14
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__24_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:41f99e0a7e90f7055be0fecc9dc4f741e880f69b421f2e5437ff7bd8fd180086
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__25_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:67bd268e7eff07e70c9baf5e936ba36a62664e93bf0633401bfa2ba0f8da3723
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__26_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:a75ef82c7afea4bd47269b140bc69e7d2bef5c567b89e7ac9b8da27c602d0ac1
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__27_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:50a18b7cc5a6138800a4ebfca5f9f7b5f113b755efdcd5bb44607a1f9ca26fbc
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__28_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:404fbc21f87420bd4986fe9d7dac743c396de45915462ade142907818a0300e5
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__29_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f8dc80b0b2152268b09070555059e4f2b7a243a32970acddd42d31af210baa06
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__2_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d0fa51aaadf768bd5f4fd6e0e034dc2b3b71509443e816137718374c4a444194
|
| 3 |
+
size 124817526
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__30_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6945798dcd184f08353f44d751abba207e473c6465fc0a80f417ac597069de8e
|
| 3 |
+
size 124815949
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__31_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cc559d42f354792db03d2e93775fc459d56bc652e7ee1e555fcf0ac49261ee53
|
| 3 |
+
size 110009933
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__3_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e22acbf9aecb9b2a4f0df58da72ae50b6c548fbfc6c848a400d8d1660d1a9bc6
|
| 3 |
+
size 124817526
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__4_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9aeb5b90f23fba031069ef1f838883ee0b811189cd227312b65b956822007664
|
| 3 |
+
size 124817526
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__5_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:89493fadeb52332e01b3d812d4d6223aaaf03ff205e967133425ba4b2d3d4bff
|
| 3 |
+
size 124817526
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__6_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3319dd99e280b5ab333b2fa14dcd76c161921a6787613971fba5b4b18f21fcf6
|
| 3 |
+
size 124817526
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__7_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:355a810253732feda0451d8f70baca4b072b72b11c6f578e69c5c64c02d1dad8
|
| 3 |
+
size 124817210
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__8_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:172e0608393f7995d4f79a283d54aa13b0c5798c88a243971d8e165f910fcccf
|
| 3 |
+
size 124817210
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/__9_0.distcp
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f2372e75c95e541de99a7eec10eb1354da3f9824a5a1bcf5b35158727ad1b875
|
| 3 |
+
size 124817210
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/checkpoint_inventory.json
ADDED
|
@@ -0,0 +1,26 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"format": "pytorch_distributed_checkpoint_fsdp2_native",
|
| 3 |
+
"source_world_size": 32,
|
| 4 |
+
"epoch": 1,
|
| 5 |
+
"step": 114274,
|
| 6 |
+
"state_inventory": {
|
| 7 |
+
"model": {
|
| 8 |
+
"tensors": 161,
|
| 9 |
+
"elements": 246673200
|
| 10 |
+
},
|
| 11 |
+
"ema": {
|
| 12 |
+
"tensors": 160,
|
| 13 |
+
"elements": 246672440
|
| 14 |
+
},
|
| 15 |
+
"exp_avg": {
|
| 16 |
+
"tensors": 160,
|
| 17 |
+
"elements": 246672440
|
| 18 |
+
},
|
| 19 |
+
"exp_avg_sq": {
|
| 20 |
+
"tensors": 160,
|
| 21 |
+
"elements": 246672440
|
| 22 |
+
}
|
| 23 |
+
},
|
| 24 |
+
"ema_location": "optim.state.<parameter>.param_ema",
|
| 25 |
+
"note": "Original model, AdamATan2 moments/counters and EMA. Trainer saved no standalone RNG/scheduler state."
|
| 26 |
+
}
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/full_checkpoint/receipt.json
ADDED
|
@@ -0,0 +1,74 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 1,
|
| 3 |
+
"epoch_step": 114274,
|
| 4 |
+
"epoch_micro_step": 114274,
|
| 5 |
+
"accumulation_steps": 1,
|
| 6 |
+
"reference_replicas": 16,
|
| 7 |
+
"gradient_reduction": "sum",
|
| 8 |
+
"gradient_reference_replicas": 16,
|
| 9 |
+
"gradient_scale": 1.0,
|
| 10 |
+
"step": 114274,
|
| 11 |
+
"input_tokens": 14922536242,
|
| 12 |
+
"response_labels": 6402595422,
|
| 13 |
+
"examples": 57699115,
|
| 14 |
+
"debug_partial_epoch": false,
|
| 15 |
+
"max_grad_norm": null,
|
| 16 |
+
"continuation": null,
|
| 17 |
+
"architecture": {
|
| 18 |
+
"arm": "lookahead",
|
| 19 |
+
"recipe_profile": "small_search_1epoch",
|
| 20 |
+
"forward_blocks": 402,
|
| 21 |
+
"hrm": null,
|
| 22 |
+
"lookahead_shape": {
|
| 23 |
+
"hidden_size": 760,
|
| 24 |
+
"intermediate_size": 2048,
|
| 25 |
+
"num_attention_heads": 10,
|
| 26 |
+
"num_hidden_layers": 16,
|
| 27 |
+
"num_passes": 3,
|
| 28 |
+
"num_refiner_layers": 3,
|
| 29 |
+
"backbone_layout": "hrm_cycle",
|
| 30 |
+
"pseudo_low_source": "real_final",
|
| 31 |
+
"l_layers": 8,
|
| 32 |
+
"l_cycles": 3,
|
| 33 |
+
"h_layers": 8,
|
| 34 |
+
"h_cycles": 2,
|
| 35 |
+
"refiner_repeats": 2,
|
| 36 |
+
"refiner_variant": "R0",
|
| 37 |
+
"latent_horizon": 2,
|
| 38 |
+
"thinking_positions": 1,
|
| 39 |
+
"thinking_selection": "first_response",
|
| 40 |
+
"refiner_gate_bias": -1.0,
|
| 41 |
+
"pseudo_candidate_mode": "soft_embedding",
|
| 42 |
+
"pseudo_top_k": 128,
|
| 43 |
+
"feedback_source": "refined_pseudo",
|
| 44 |
+
"refiner_memory": true,
|
| 45 |
+
"refiner_memory_only": true,
|
| 46 |
+
"cache_layer_indices": null
|
| 47 |
+
},
|
| 48 |
+
"response_ce_weights": [
|
| 49 |
+
0.5,
|
| 50 |
+
0.25,
|
| 51 |
+
0.25
|
| 52 |
+
],
|
| 53 |
+
"auxiliary_losses": {
|
| 54 |
+
"draft_loss_weight": 0,
|
| 55 |
+
"future_loss_weight": 0.2,
|
| 56 |
+
"refined_future_loss_weight": 0.0,
|
| 57 |
+
"auxiliary_loss_passes": "all",
|
| 58 |
+
"auxiliary_loss_weighting": "fbw",
|
| 59 |
+
"future_gradient_mode": "all",
|
| 60 |
+
"normalization": "divide primary + weighted D/P by 1 for K1, 2 for K>=2"
|
| 61 |
+
},
|
| 62 |
+
"pass_schedule": null,
|
| 63 |
+
"forward_blocks_convention": "maximum K; actual K and block calls logged each update"
|
| 64 |
+
},
|
| 65 |
+
"pass_schedule": null,
|
| 66 |
+
"pass_accounting": {
|
| 67 |
+
"3": {
|
| 68 |
+
"updates": 114274,
|
| 69 |
+
"input_tokens": 14922536242,
|
| 70 |
+
"response_labels": 6402595422,
|
| 71 |
+
"examples": 57699115
|
| 72 |
+
}
|
| 73 |
+
}
|
| 74 |
+
}
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/manifest.json
ADDED
|
@@ -0,0 +1,191 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"schema_version": 1,
|
| 3 |
+
"repo_id": "Siyuc/loopTF-workspace",
|
| 4 |
+
"path_in_repo": "small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274",
|
| 5 |
+
"files": {
|
| 6 |
+
"README.md": {
|
| 7 |
+
"size": 1238,
|
| 8 |
+
"sha256": "b61cce4679bdec8b984622aef2945fd29a4748305aa4ab051c4f87a6b142ce08"
|
| 9 |
+
},
|
| 10 |
+
"code.diff": {
|
| 11 |
+
"size": 0,
|
| 12 |
+
"sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
|
| 13 |
+
},
|
| 14 |
+
"complete.json": {
|
| 15 |
+
"size": 792,
|
| 16 |
+
"sha256": "7dcc9245179a47b2085773af80108ddf2ed17e1d92844fe39765f8af8c4b1e76"
|
| 17 |
+
},
|
| 18 |
+
"config.json": {
|
| 19 |
+
"size": 3833,
|
| 20 |
+
"sha256": "560d5405b708e4e0b2d775b7fa870438fdbf003f65ef9086b7a49bb999108dff"
|
| 21 |
+
},
|
| 22 |
+
"config.yaml": {
|
| 23 |
+
"size": 1729,
|
| 24 |
+
"sha256": "e4d47ba4c356b1be78e8ffd729e468768713e32278a4e691c3b1b936c1620f47"
|
| 25 |
+
},
|
| 26 |
+
"full_checkpoint/.metadata": {
|
| 27 |
+
"size": 3914233,
|
| 28 |
+
"sha256": "b97f4d041169fa503d9cdf05793d8577f82d990add2c4a9d2c2a118bd0bc9451"
|
| 29 |
+
},
|
| 30 |
+
"full_checkpoint/__0_0.distcp": {
|
| 31 |
+
"size": 124811113,
|
| 32 |
+
"sha256": "98c78e37ea6c2c9927517422b826d672b78f821b5a9bfc998591245b5e50e424"
|
| 33 |
+
},
|
| 34 |
+
"full_checkpoint/__10_0.distcp": {
|
| 35 |
+
"size": 124824058,
|
| 36 |
+
"sha256": "7fbb54c0bfdec89cec63946da7cce5c710c727cdf187319e096303046ff26b17"
|
| 37 |
+
},
|
| 38 |
+
"full_checkpoint/__11_0.distcp": {
|
| 39 |
+
"size": 124815949,
|
| 40 |
+
"sha256": "386ea6f23ae99b33689d6ce799a749d5cd9247857889a862c3365c1079e2fc12"
|
| 41 |
+
},
|
| 42 |
+
"full_checkpoint/__12_0.distcp": {
|
| 43 |
+
"size": 124815949,
|
| 44 |
+
"sha256": "577f7c603621263cd49e94e76bbaf23f10a14b56643d1e6a2f109fc251b8ca04"
|
| 45 |
+
},
|
| 46 |
+
"full_checkpoint/__13_0.distcp": {
|
| 47 |
+
"size": 124815949,
|
| 48 |
+
"sha256": "a19f429adb42c7a147c086d697c1fe489590f721ef706b2d95abcb33e2fccd10"
|
| 49 |
+
},
|
| 50 |
+
"full_checkpoint/__14_0.distcp": {
|
| 51 |
+
"size": 124815949,
|
| 52 |
+
"sha256": "13ebd9be44bb65dc2bc7ef3c6fcae44cf53b59efe101b49ef4756651842ce032"
|
| 53 |
+
},
|
| 54 |
+
"full_checkpoint/__15_0.distcp": {
|
| 55 |
+
"size": 124815949,
|
| 56 |
+
"sha256": "1028ad750e6eb5cd1c4ce4de9e85897ab16da4199c8852c2ca9ff4ffcbecd2b4"
|
| 57 |
+
},
|
| 58 |
+
"full_checkpoint/__16_0.distcp": {
|
| 59 |
+
"size": 124815949,
|
| 60 |
+
"sha256": "a037c8e382b239d3a1a87666ca4892cdaffecb9b35a5dddd2899c4373d9523db"
|
| 61 |
+
},
|
| 62 |
+
"full_checkpoint/__17_0.distcp": {
|
| 63 |
+
"size": 124815949,
|
| 64 |
+
"sha256": "8ea6d72044c2fb61d26812001fd78803a4e707079e217ef2a63119962261678a"
|
| 65 |
+
},
|
| 66 |
+
"full_checkpoint/__18_0.distcp": {
|
| 67 |
+
"size": 124815949,
|
| 68 |
+
"sha256": "8ff03bbc6d0a974dfc75b2b14c3a305e8851536172bb158fd63c794c73a88637"
|
| 69 |
+
},
|
| 70 |
+
"full_checkpoint/__19_0.distcp": {
|
| 71 |
+
"size": 124815949,
|
| 72 |
+
"sha256": "031e6bc9958fa694f9249d86497b398331044bb114eff9abeee7afb1ec31bd06"
|
| 73 |
+
},
|
| 74 |
+
"full_checkpoint/__1_0.distcp": {
|
| 75 |
+
"size": 124817526,
|
| 76 |
+
"sha256": "daf585fcfa65f3927e8c02a0c7ab5f8e46a9ba280d61e3b959b6bdd130973a49"
|
| 77 |
+
},
|
| 78 |
+
"full_checkpoint/__20_0.distcp": {
|
| 79 |
+
"size": 124815949,
|
| 80 |
+
"sha256": "56e7c51587ff1bde1cfbc3e2492dc28216199ef9f57a0ac0b2c66f6fdd875d85"
|
| 81 |
+
},
|
| 82 |
+
"full_checkpoint/__21_0.distcp": {
|
| 83 |
+
"size": 124815949,
|
| 84 |
+
"sha256": "e52b270e8abb3858fd676c352120772401e8298a53db680c438a548f302e74ab"
|
| 85 |
+
},
|
| 86 |
+
"full_checkpoint/__22_0.distcp": {
|
| 87 |
+
"size": 124815949,
|
| 88 |
+
"sha256": "e44acb5b968124b89fc6011448a7bc5e8b2b92a6cf71c28d26b373ded032cd5b"
|
| 89 |
+
},
|
| 90 |
+
"full_checkpoint/__23_0.distcp": {
|
| 91 |
+
"size": 124815949,
|
| 92 |
+
"sha256": "f1474e6bf7cd4209a951e44d93ee0014eef4ea988b476612438fed67d8270b14"
|
| 93 |
+
},
|
| 94 |
+
"full_checkpoint/__24_0.distcp": {
|
| 95 |
+
"size": 124815949,
|
| 96 |
+
"sha256": "41f99e0a7e90f7055be0fecc9dc4f741e880f69b421f2e5437ff7bd8fd180086"
|
| 97 |
+
},
|
| 98 |
+
"full_checkpoint/__25_0.distcp": {
|
| 99 |
+
"size": 124815949,
|
| 100 |
+
"sha256": "67bd268e7eff07e70c9baf5e936ba36a62664e93bf0633401bfa2ba0f8da3723"
|
| 101 |
+
},
|
| 102 |
+
"full_checkpoint/__26_0.distcp": {
|
| 103 |
+
"size": 124815949,
|
| 104 |
+
"sha256": "a75ef82c7afea4bd47269b140bc69e7d2bef5c567b89e7ac9b8da27c602d0ac1"
|
| 105 |
+
},
|
| 106 |
+
"full_checkpoint/__27_0.distcp": {
|
| 107 |
+
"size": 124815949,
|
| 108 |
+
"sha256": "50a18b7cc5a6138800a4ebfca5f9f7b5f113b755efdcd5bb44607a1f9ca26fbc"
|
| 109 |
+
},
|
| 110 |
+
"full_checkpoint/__28_0.distcp": {
|
| 111 |
+
"size": 124815949,
|
| 112 |
+
"sha256": "404fbc21f87420bd4986fe9d7dac743c396de45915462ade142907818a0300e5"
|
| 113 |
+
},
|
| 114 |
+
"full_checkpoint/__29_0.distcp": {
|
| 115 |
+
"size": 124815949,
|
| 116 |
+
"sha256": "f8dc80b0b2152268b09070555059e4f2b7a243a32970acddd42d31af210baa06"
|
| 117 |
+
},
|
| 118 |
+
"full_checkpoint/__2_0.distcp": {
|
| 119 |
+
"size": 124817526,
|
| 120 |
+
"sha256": "d0fa51aaadf768bd5f4fd6e0e034dc2b3b71509443e816137718374c4a444194"
|
| 121 |
+
},
|
| 122 |
+
"full_checkpoint/__30_0.distcp": {
|
| 123 |
+
"size": 124815949,
|
| 124 |
+
"sha256": "6945798dcd184f08353f44d751abba207e473c6465fc0a80f417ac597069de8e"
|
| 125 |
+
},
|
| 126 |
+
"full_checkpoint/__31_0.distcp": {
|
| 127 |
+
"size": 110009933,
|
| 128 |
+
"sha256": "cc559d42f354792db03d2e93775fc459d56bc652e7ee1e555fcf0ac49261ee53"
|
| 129 |
+
},
|
| 130 |
+
"full_checkpoint/__3_0.distcp": {
|
| 131 |
+
"size": 124817526,
|
| 132 |
+
"sha256": "e22acbf9aecb9b2a4f0df58da72ae50b6c548fbfc6c848a400d8d1660d1a9bc6"
|
| 133 |
+
},
|
| 134 |
+
"full_checkpoint/__4_0.distcp": {
|
| 135 |
+
"size": 124817526,
|
| 136 |
+
"sha256": "9aeb5b90f23fba031069ef1f838883ee0b811189cd227312b65b956822007664"
|
| 137 |
+
},
|
| 138 |
+
"full_checkpoint/__5_0.distcp": {
|
| 139 |
+
"size": 124817526,
|
| 140 |
+
"sha256": "89493fadeb52332e01b3d812d4d6223aaaf03ff205e967133425ba4b2d3d4bff"
|
| 141 |
+
},
|
| 142 |
+
"full_checkpoint/__6_0.distcp": {
|
| 143 |
+
"size": 124817526,
|
| 144 |
+
"sha256": "3319dd99e280b5ab333b2fa14dcd76c161921a6787613971fba5b4b18f21fcf6"
|
| 145 |
+
},
|
| 146 |
+
"full_checkpoint/__7_0.distcp": {
|
| 147 |
+
"size": 124817210,
|
| 148 |
+
"sha256": "355a810253732feda0451d8f70baca4b072b72b11c6f578e69c5c64c02d1dad8"
|
| 149 |
+
},
|
| 150 |
+
"full_checkpoint/__8_0.distcp": {
|
| 151 |
+
"size": 124817210,
|
| 152 |
+
"sha256": "172e0608393f7995d4f79a283d54aa13b0c5798c88a243971d8e165f910fcccf"
|
| 153 |
+
},
|
| 154 |
+
"full_checkpoint/__9_0.distcp": {
|
| 155 |
+
"size": 124817210,
|
| 156 |
+
"sha256": "f2372e75c95e541de99a7eec10eb1354da3f9824a5a1bcf5b35158727ad1b875"
|
| 157 |
+
},
|
| 158 |
+
"full_checkpoint/checkpoint_inventory.json": {
|
| 159 |
+
"size": 610,
|
| 160 |
+
"sha256": "9176fe623fd875c9f04aacd6cdb2ddcb4f1b26fd3520732d06fb0567f1fc4e52"
|
| 161 |
+
},
|
| 162 |
+
"full_checkpoint/receipt.json": {
|
| 163 |
+
"size": 2043,
|
| 164 |
+
"sha256": "309571c204a672532c15b8af78e571c0471bb4ddad49990fa3e3a049bbea07fc"
|
| 165 |
+
},
|
| 166 |
+
"model.safetensors": {
|
| 167 |
+
"size": 986709008,
|
| 168 |
+
"sha256": "705f6961462add6ac501cebe28b7e45c8ebba200a6b955d3594fa386a5ce4b66"
|
| 169 |
+
},
|
| 170 |
+
"original_model.safetensors": {
|
| 171 |
+
"size": 986709008,
|
| 172 |
+
"sha256": "ddba39a533f5d29943339fb28417068e0ae44ec1844a3cc4c7c938341fc674f8"
|
| 173 |
+
},
|
| 174 |
+
"recipe.json": {
|
| 175 |
+
"size": 9144,
|
| 176 |
+
"sha256": "e47d5b7b800e9ff38f8a662649b1b9e6431ffb050f43e3f7246f5a97b3feefc4"
|
| 177 |
+
},
|
| 178 |
+
"snapshot_manifest.json": {
|
| 179 |
+
"size": 66253,
|
| 180 |
+
"sha256": "fcd59deb66130782e1a3ad55e545f3444d8b587e5e8415b5600e9629b4e66c95"
|
| 181 |
+
},
|
| 182 |
+
"tokenizer/tokenizer.json": {
|
| 183 |
+
"size": 4732721,
|
| 184 |
+
"sha256": "d569c356b9c88d4fb5fc2ab43f0b9f9c445336d3121588f1f2f6e2bdbc6fdc0f"
|
| 185 |
+
},
|
| 186 |
+
"training_code.tar.gz": {
|
| 187 |
+
"size": 626824,
|
| 188 |
+
"sha256": "e23bf06d79b84d6b355509b3e857928ce842852419f1c1a2a39a6e0ea37669c1"
|
| 189 |
+
}
|
| 190 |
+
}
|
| 191 |
+
}
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:705f6961462add6ac501cebe28b7e45c8ebba200a6b955d3594fa386a5ce4b66
|
| 3 |
+
size 986709008
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/original_model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ddba39a533f5d29943339fb28417068e0ae44ec1844a3cc4c7c938341fc674f8
|
| 3 |
+
size 986709008
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/recipe.json
ADDED
|
@@ -0,0 +1,278 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"launching_card": "/root/dl/runtime/script/loopTF-workspace/launchings/product/20261010_hrm_document_redistribution/20261010-034356_bins16-dp32.md",
|
| 3 |
+
"upstream_revision": "aaa948ea674fd84b7bc455c9cfb455ecfefdf914",
|
| 4 |
+
"recipe": {
|
| 5 |
+
"arch": {
|
| 6 |
+
"name": "baselines.hrm_nocarry_bp_warmup@HierarchicalReasoningModel",
|
| 7 |
+
"head": "lm_head@LMHead",
|
| 8 |
+
"half_layers": false,
|
| 9 |
+
"H_cycles": 2,
|
| 10 |
+
"L_cycles": 3,
|
| 11 |
+
"H_override": {
|
| 12 |
+
"n_layers": 8
|
| 13 |
+
},
|
| 14 |
+
"bp_warmup_ratio": 0.2,
|
| 15 |
+
"bp_max_steps": 5,
|
| 16 |
+
"n_layers": 8,
|
| 17 |
+
"hidden_size": 760,
|
| 18 |
+
"num_heads": 10,
|
| 19 |
+
"expansion": 4,
|
| 20 |
+
"norm_type": "pre",
|
| 21 |
+
"norm_eps": 1e-06,
|
| 22 |
+
"rope_theta": 10000.0,
|
| 23 |
+
"pos_emb_type": "rope",
|
| 24 |
+
"init_type": "lecun_normal",
|
| 25 |
+
"attn_type": "causal"
|
| 26 |
+
},
|
| 27 |
+
"data": {
|
| 28 |
+
"path": "/root/dl/runtime/script/loopTF-workspace/data/causal_hrm_text_sources/sampled",
|
| 29 |
+
"target_only": true
|
| 30 |
+
},
|
| 31 |
+
"global_batch_size": 131072,
|
| 32 |
+
"epochs": 1,
|
| 33 |
+
"lr": 0.0003,
|
| 34 |
+
"lr_min_ratio": 1.0,
|
| 35 |
+
"lr_warmup_steps": 750,
|
| 36 |
+
"weight_decay": 0.1,
|
| 37 |
+
"beta1": 0.9,
|
| 38 |
+
"beta2": 0.95,
|
| 39 |
+
"ema": 0.9999,
|
| 40 |
+
"fwd_bwd_dtype": "bfloat16",
|
| 41 |
+
"project_name": "lookahead_hrm_small_search",
|
| 42 |
+
"run_name": "Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32",
|
| 43 |
+
"checkpoint_path": "/apdcephfs_tj31/share_303214315/hunyuan/kenanchen/loopTF-workspace/output/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32/checkpoints",
|
| 44 |
+
"resume_from": null,
|
| 45 |
+
"resume_epoch": null,
|
| 46 |
+
"weights_only_resume_from_ema": false,
|
| 47 |
+
"seed": 0,
|
| 48 |
+
"checkpoint_interval": 1,
|
| 49 |
+
"log_interval": 5
|
| 50 |
+
},
|
| 51 |
+
"parameters": 246672440,
|
| 52 |
+
"world_size": 32,
|
| 53 |
+
"local_batch_tokens": null,
|
| 54 |
+
"estimated_total_steps": 114296,
|
| 55 |
+
"execution": {
|
| 56 |
+
"stop_at_unix_time": null,
|
| 57 |
+
"runtime_cache_rank0": "/root/dl/runtime/script/loopTF-workspace/data/cache/hrm_official/runtime/57b584501d0c9deee6f2/rank0",
|
| 58 |
+
"attention": "flex_triton_segmented_causal",
|
| 59 |
+
"checkpointing": false,
|
| 60 |
+
"checkpointing_skip_every": 0,
|
| 61 |
+
"checkpointing_retain_two_thirds_k2": false,
|
| 62 |
+
"checkpointing_min_passes": 0,
|
| 63 |
+
"compile_blocks": true,
|
| 64 |
+
"compile_hrm_stacks": false,
|
| 65 |
+
"share_projection_weights": false,
|
| 66 |
+
"hrm_fsdp_grouping": "block",
|
| 67 |
+
"compile_optimizer": true,
|
| 68 |
+
"compile_vocabulary_loss": true,
|
| 69 |
+
"split_pseudo_attention": true,
|
| 70 |
+
"compile_a2_mix": true,
|
| 71 |
+
"lookahead_fsdp_grouping": "model",
|
| 72 |
+
"accumulation_steps": 1,
|
| 73 |
+
"gradient_reduction": "sum",
|
| 74 |
+
"gradient_reference_replicas": 16,
|
| 75 |
+
"gradient_scale": 1.0,
|
| 76 |
+
"virtual_replicas": 16,
|
| 77 |
+
"micro_batch_tokens": 131072,
|
| 78 |
+
"redistribution": {
|
| 79 |
+
"mode": "whole_document",
|
| 80 |
+
"reference_replicas": 16,
|
| 81 |
+
"physical_replicas": 32,
|
| 82 |
+
"expansion": 2,
|
| 83 |
+
"reference_batch_tokens": 8192,
|
| 84 |
+
"padding": "multiple_of_128",
|
| 85 |
+
"initialization": "reference_fsdp_row_shards",
|
| 86 |
+
"gradient_normalization": "global_response_count/reference_replicas",
|
| 87 |
+
"future_normalization": "reference_bin_response_count/future_count"
|
| 88 |
+
}
|
| 89 |
+
},
|
| 90 |
+
"debug": {
|
| 91 |
+
"max_steps": 0,
|
| 92 |
+
"force_bp": 0,
|
| 93 |
+
"force_passes": null,
|
| 94 |
+
"benchmark_execution": false,
|
| 95 |
+
"step_offset": 0
|
| 96 |
+
},
|
| 97 |
+
"data_order": "official multipack, no holdout, upstream per-epoch drop_last, packed for 16 virtual replicas",
|
| 98 |
+
"torch": "2.13.0+cu129",
|
| 99 |
+
"cuda": "12.9",
|
| 100 |
+
"gpu": "NVIDIA H800",
|
| 101 |
+
"config_sha256": "e4d47ba4c356b1be78e8ffd729e468768713e32278a4e691c3b1b936c1620f47",
|
| 102 |
+
"resolved_overrides": {
|
| 103 |
+
"arch": {
|
| 104 |
+
"H_cycles": 2,
|
| 105 |
+
"H_override": {
|
| 106 |
+
"n_layers": 8
|
| 107 |
+
},
|
| 108 |
+
"L_cycles": 3,
|
| 109 |
+
"attn_type": "causal",
|
| 110 |
+
"expansion": 4,
|
| 111 |
+
"half_layers": false,
|
| 112 |
+
"hidden_size": 760,
|
| 113 |
+
"n_layers": 8,
|
| 114 |
+
"num_heads": 10
|
| 115 |
+
},
|
| 116 |
+
"checkpoint_path": "/root/dl/runtime/script/loopTF-workspace/output/causal_hrm_text/small_search/double_compute/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch",
|
| 117 |
+
"data": {
|
| 118 |
+
"path": "/root/dl/runtime/script/loopTF-workspace/data/causal_hrm_text_sources/sampled"
|
| 119 |
+
},
|
| 120 |
+
"epochs": 1,
|
| 121 |
+
"global_batch_size": 131072,
|
| 122 |
+
"lr": 0.0003,
|
| 123 |
+
"lr_warmup_steps": 750,
|
| 124 |
+
"project_name": "lookahead_hrm_small_search",
|
| 125 |
+
"run_name": "Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch"
|
| 126 |
+
},
|
| 127 |
+
"continuation": null,
|
| 128 |
+
"resume": null,
|
| 129 |
+
"resume_execution": null,
|
| 130 |
+
"checkpointing_interval": {
|
| 131 |
+
"save_every": 1000,
|
| 132 |
+
"keep_step_checkpoints": 2,
|
| 133 |
+
"save_first_step": true,
|
| 134 |
+
"verify_first": true
|
| 135 |
+
},
|
| 136 |
+
"max_grad_norm": null,
|
| 137 |
+
"pass_schedule": null,
|
| 138 |
+
"reference_validation": {
|
| 139 |
+
"suite": "/apdcephfs_tj31/share_303214315/hunyuan/kenanchen/loopTF-workspace/data/evaluation/hrm_text/reference-v1-seed20261004",
|
| 140 |
+
"manifest_sha256": "98c9561e9ae9ebf972500e8dea63dde867b356f6e85016dd4a950120cc2635ba",
|
| 141 |
+
"weights": "ema",
|
| 142 |
+
"capacity": 4096,
|
| 143 |
+
"examples": 1280,
|
| 144 |
+
"teacher_forced": true,
|
| 145 |
+
"tasks": {
|
| 146 |
+
"GSM8k": {
|
| 147 |
+
"examples": 256,
|
| 148 |
+
"response_tokens": 24971,
|
| 149 |
+
"packs": 11
|
| 150 |
+
},
|
| 151 |
+
"MATH": {
|
| 152 |
+
"examples": 256,
|
| 153 |
+
"response_tokens": 56990,
|
| 154 |
+
"packs": 21
|
| 155 |
+
},
|
| 156 |
+
"DROP": {
|
| 157 |
+
"examples": 128,
|
| 158 |
+
"response_tokens": 533,
|
| 159 |
+
"packs": 54
|
| 160 |
+
},
|
| 161 |
+
"MMLU": {
|
| 162 |
+
"examples": 128,
|
| 163 |
+
"response_tokens": 128,
|
| 164 |
+
"packs": 22
|
| 165 |
+
},
|
| 166 |
+
"ARC": {
|
| 167 |
+
"examples": 128,
|
| 168 |
+
"response_tokens": 128,
|
| 169 |
+
"packs": 128
|
| 170 |
+
},
|
| 171 |
+
"HellaSwag": {
|
| 172 |
+
"examples": 128,
|
| 173 |
+
"response_tokens": 128,
|
| 174 |
+
"packs": 64
|
| 175 |
+
},
|
| 176 |
+
"Winogrande": {
|
| 177 |
+
"examples": 128,
|
| 178 |
+
"response_tokens": 128,
|
| 179 |
+
"packs": 7
|
| 180 |
+
},
|
| 181 |
+
"BoolQ": {
|
| 182 |
+
"examples": 128,
|
| 183 |
+
"response_tokens": 128,
|
| 184 |
+
"packs": 39
|
| 185 |
+
}
|
| 186 |
+
},
|
| 187 |
+
"every": 1000,
|
| 188 |
+
"first_update": true,
|
| 189 |
+
"training_horizon_changed": false
|
| 190 |
+
},
|
| 191 |
+
"tracking": {
|
| 192 |
+
"mode": "online",
|
| 193 |
+
"entity": "LoopTF-4-CSPs",
|
| 194 |
+
"project": "lookahead_hrm_small_search",
|
| 195 |
+
"group": "small_search_1epoch"
|
| 196 |
+
},
|
| 197 |
+
"architecture": {
|
| 198 |
+
"arm": "lookahead",
|
| 199 |
+
"recipe_profile": "small_search_1epoch",
|
| 200 |
+
"forward_blocks": 402,
|
| 201 |
+
"hrm": null,
|
| 202 |
+
"lookahead_shape": {
|
| 203 |
+
"hidden_size": 760,
|
| 204 |
+
"intermediate_size": 2048,
|
| 205 |
+
"num_attention_heads": 10,
|
| 206 |
+
"num_hidden_layers": 16,
|
| 207 |
+
"num_passes": 3,
|
| 208 |
+
"num_refiner_layers": 3,
|
| 209 |
+
"backbone_layout": "hrm_cycle",
|
| 210 |
+
"pseudo_low_source": "real_final",
|
| 211 |
+
"l_layers": 8,
|
| 212 |
+
"l_cycles": 3,
|
| 213 |
+
"h_layers": 8,
|
| 214 |
+
"h_cycles": 2,
|
| 215 |
+
"refiner_repeats": 2,
|
| 216 |
+
"refiner_variant": "R0",
|
| 217 |
+
"latent_horizon": 2,
|
| 218 |
+
"thinking_positions": 1,
|
| 219 |
+
"thinking_selection": "first_response",
|
| 220 |
+
"refiner_gate_bias": -1.0,
|
| 221 |
+
"pseudo_candidate_mode": "soft_embedding",
|
| 222 |
+
"pseudo_top_k": 128,
|
| 223 |
+
"feedback_source": "refined_pseudo",
|
| 224 |
+
"refiner_memory": true,
|
| 225 |
+
"refiner_memory_only": true,
|
| 226 |
+
"cache_layer_indices": null
|
| 227 |
+
},
|
| 228 |
+
"response_ce_weights": [
|
| 229 |
+
0.5,
|
| 230 |
+
0.25,
|
| 231 |
+
0.25
|
| 232 |
+
],
|
| 233 |
+
"auxiliary_losses": {
|
| 234 |
+
"draft_loss_weight": 0,
|
| 235 |
+
"future_loss_weight": 0.2,
|
| 236 |
+
"refined_future_loss_weight": 0.0,
|
| 237 |
+
"auxiliary_loss_passes": "all",
|
| 238 |
+
"auxiliary_loss_weighting": "fbw",
|
| 239 |
+
"future_gradient_mode": "all",
|
| 240 |
+
"normalization": "divide primary + weighted D/P by 1 for K1, 2 for K>=2"
|
| 241 |
+
},
|
| 242 |
+
"pass_schedule": null,
|
| 243 |
+
"forward_blocks_convention": "maximum K; actual K and block calls logged each update",
|
| 244 |
+
"lookahead": {
|
| 245 |
+
"mode": "parallel",
|
| 246 |
+
"num_passes": 3,
|
| 247 |
+
"feedback_loss_weight": 1.0,
|
| 248 |
+
"attention_backend": "triton",
|
| 249 |
+
"checkpoint_refiner": false,
|
| 250 |
+
"cache_layer_indices": null,
|
| 251 |
+
"num_refiner_layers": 3,
|
| 252 |
+
"refiner_activation": "swiglu",
|
| 253 |
+
"refiner_intermediate_size": null,
|
| 254 |
+
"refiner_gate_bias": -1.0,
|
| 255 |
+
"future_loss_weight": 0.2,
|
| 256 |
+
"future_gradient_mode": "all",
|
| 257 |
+
"draft_loss_weight": 0,
|
| 258 |
+
"auxiliary_loss_passes": "all",
|
| 259 |
+
"auxiliary_loss_weighting": "fbw",
|
| 260 |
+
"refined_future_loss_weight": 0.0,
|
| 261 |
+
"pseudo_initializer": "A2",
|
| 262 |
+
"pseudo_candidate_mode": "soft_embedding",
|
| 263 |
+
"pseudo_top_k": 128,
|
| 264 |
+
"pseudo_temperature": 1.0,
|
| 265 |
+
"normalize_feedback_embedding": false,
|
| 266 |
+
"normalize_feedback_output": false,
|
| 267 |
+
"feedback_source": "refined_pseudo",
|
| 268 |
+
"refiner_variant": "R0",
|
| 269 |
+
"latent_horizon": 2,
|
| 270 |
+
"thinking_positions": 1,
|
| 271 |
+
"thinking_selection": "first_response",
|
| 272 |
+
"refiner_memory": true,
|
| 273 |
+
"refiner_memory_only": true,
|
| 274 |
+
"allow_objective_change_on_resume": false
|
| 275 |
+
}
|
| 276 |
+
},
|
| 277 |
+
"objective": "normalized_feedback_pass_response_CE_with_final_pass_draft_and_future_CE"
|
| 278 |
+
}
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/snapshot_manifest.json
ADDED
|
@@ -0,0 +1,497 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"config.yaml": "e4d47ba4c356b1be78e8ffd729e468768713e32278a4e691c3b1b936c1620f47",
|
| 3 |
+
"configs/README.md": "d27e6a99230d94161e440d9c60c79334e30a310ebb1d7e2f66d52f2aa6582d18",
|
| 4 |
+
"configs/benchmarks/done_right_pretrain_v3.json": "0ca58a6d7757e31ff5fd72c95c4f350f61a3cdd88a3f8f66b6554f7eec512c6b",
|
| 5 |
+
"configs/benchmarks/hrm_text_extended_v1.json": "6654e1661943fabd6e577e7be206975b2e78d383b2bda3351103ebe06267ea07",
|
| 6 |
+
"configs/benchmarks/hrm_text_official_v1.json": "e01e8a8faaa199daaf16a32eb30cc704c035f4522531db6200e3c09d30dfd52f",
|
| 7 |
+
"configs/benchmarks/ifm_v1.json": "9542228f1a844167f502b225d699451d4b43a97c148af5c19709c3150c0ebf7b",
|
| 8 |
+
"configs/benchmarks/paper2604_prefix_query_v1.json": "b1c7bec8c8df3955fa075e82ce6f9647c8858a5605d6833628c167d209a314c4",
|
| 9 |
+
"configs/benchmarks/paper2604_v1.json": "22f9cb71bf847933136d65dbd25d1659b2b6c7bc53e025d043edbede94872bc4",
|
| 10 |
+
"configs/causal_hrm_text/HRM-H2-L3-H18-L15-BP2to5-seed42.yaml": "9b332109632ab89a958b22e1f3f64a83df865d5cef4f43972322fb2e4c4b682f",
|
| 11 |
+
"configs/causal_hrm_text/Lookahead-K2-LA2-A2-R0-D0-P0-F0-seed42-B30-Ref3-hrmBlock-refinerMemoryOnly.yaml": "d7e18550f13fc07864cf3f0dcecc83eb9989faa287ba21fa5a52a4df2d8c3668",
|
| 12 |
+
"configs/causal_hrm_text/Lookahead-K3-LA2-A2-R0-D0-P0-F0-seed42-B19-Ref4-H1904-FFN4096-hrmBlock-refinerMemoryOnly.yaml": "812d6429764fb57b6e37a1dbe0fafdf6287b8d10eccecee2ea0ea016bbbba861",
|
| 13 |
+
"configs/causal_hrm_text/Lookahead-K3-LA2-A2-R0-D0-P0-F0-seed42-B30-Ref3-hrmBlock-refinerMemoryOnly.yaml": "a066ba04d8123f0a3728707f7c4d2b00d2f848bfd02274ef38dc07afef0c2c47",
|
| 14 |
+
"configs/causal_hrm_text/README.md": "e8fe618066a1dadda99f04787eaa7f128edf3d19038b84bf0ae77172552e085f",
|
| 15 |
+
"configs/causal_hrm_text/common.yaml": "793330f1df1ef7a379c37026a27d546cf99075260d46e6cb3fa91c79a8e3e631",
|
| 16 |
+
"configs/causal_hrm_text/manifest.example.json": "15b773b642744cb00ff5e9c149362b1da5a9fb7057711e92035b1951727fcc64",
|
| 17 |
+
"configs/causal_hrm_text/official_recipe/HRM-H2-L3-H18-L15-causal-AdamATan2-EMA-seed0.yaml": "35fd40b9824b3e5ba1e28521e797a5b6a492fd24d43111bb7de02cd5c4b91804",
|
| 18 |
+
"configs/causal_hrm_text/official_recipe/LOOKAHEAD_PROGRESSIVE_DESIGN.md": "593ff3435c3469391db440d6497cce818927350c5680608a7505918c14763f3e",
|
| 19 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K2-LA2-A2-R0-D0-P0-F0-seed0-B30-Ref3-AdamATan2-EMA.yaml": "611b403c8105497b9b700538b9a544d63b321f09603e56c2cb301b19213f3082",
|
| 20 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K2-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA.yaml": "6d824820ee7c332c20f9298707d0a3bca8f78deb9bbbd2beca30d80a7cece6df",
|
| 21 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0-P0-F0-seed0-B19-Ref4-H1904-FFN4096-hrmBlock-refinerMemoryOnly-AdamATan2-EMA.yaml": "dddcb67f5fc2c3c9a0b17f1622a495577de0a68387a9280d4fe42db8f55d469e",
|
| 22 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0-P0-F0-seed0-B30-Ref3-AdamATan2-EMA-progressiveK-75-90.yaml": "0ea49aeeab27a0d4c163df88daca698be6a1e322a6ca13e61aa1218dbd587bea",
|
| 23 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0-P0-F0-seed0-B30-Ref3-AdamATan2-EMA.yaml": "f54f290a3df7a122d0326d272de2e614eca0d0f6bf34a099fc6c242387a9afc8",
|
| 24 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B19-Ref4-H1904-FFN4096-hrmBlock-refinerMemoryOnly-AdamATan2-EMA.yaml": "2ca84993c28ece114c87899ca69ea694d41d5822e2dfd6f4920d3dea016493b2",
|
| 25 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA-progressiveK-75-90.yaml": "a4a880078167ce52c844e909ca337148da2e51a0177c905f4486c8a9e554d989",
|
| 26 |
+
"configs/causal_hrm_text/official_recipe/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-B30-Ref3-AdamATan2-EMA.yaml": "ec2cb16e2f138f6e119c60c6298b8b76b59bd6efbacc74e2c2a7cb1a0e03518d",
|
| 27 |
+
"configs/causal_hrm_text/official_recipe/README.md": "792e150e1663845754efbf2628980796d6fe457edbf621fe86955754c68f5f06",
|
| 28 |
+
"configs/causal_hrm_text/official_recipe/base/lookahead_hrm_cycle.yaml": "b050ddc1dc4698b650fb7141953a6e682488c9d87fa516e550f8b0bba135452d",
|
| 29 |
+
"configs/causal_hrm_text/official_recipe/base/small_hrm.yaml": "6e8ad749700873616eebec01a7b411a7e48330c01cb39538457ca7741bf4550f",
|
| 30 |
+
"configs/causal_hrm_text/official_recipe/base/small_lookahead_hrm_cycle.yaml": "8e99a54652bc55b16f5fd6d76b218bc82c48994ad89104212ea632be5f821402",
|
| 31 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K1-LA2-A2-R0-D0-P0-F0-seed0-H18-L15x3-hrmCycle-Ref3-AdamATan2-EMA.yaml": "08b976173d53cdeed0ea4c5ade56319b74d3da95ab6bc04980935653d2c07b3d",
|
| 32 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K1-LA2-A2-R0-D0-P0p2-F0-seed0-W1440-H18-L15x3-hrmCycle-Ref3-lowRealFinal-AdamATan2-EMA.yaml": "fedb1c5fac44b7898d4414eff4358654c270171264c7e36475ec5620a4cca0ab",
|
| 33 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-H18-L15x3-hrmCycle-Ref3-AdamATan2-EMA.yaml": "add4cda7c47d6574aecbd43322654971d5d4db79c1950e32a8738862b06529e7",
|
| 34 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H15-L15x3-hrmCycle-Ref6-lowRealFinal-AdamATan2-EMA.yaml": "cb24da66893f8e4bcae3261e11c4d7bcf9b39f61889138dceb756d0806cfe79e",
|
| 35 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H18-L15x3-hrmCycle-Ref3-lowRealFinal-AdamATan2-EMA-fixed.yaml": "3700dd72eda549d952927aac5567846a8f9dc74c9058d1f2c38e0348c6e04391",
|
| 36 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H18-L15x3-hrmCycle-Ref3-lowRealFinal-AdamATan2-EMA.yaml": "b3f4fa07f7cbc69c0777652c74d92cdf35053cc20afb7606de704039515d69f1",
|
| 37 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-H18-L15x3-hrmCycle-Ref3-AdamATan2-EMA-progressiveK-75-90.yaml": "851d10bf54b8ad9199c21108eb4d879754b1921c812ee6eaa1709eb47e24b0d2",
|
| 38 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed0-H18-L15x3-hrmCycle-Ref3-AdamATan2-EMA.yaml": "07f11242383826d5fd8af531e95a3d90baa7fa06a2c8766adb45767e544c73a5",
|
| 39 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/README.md": "4f21d5b172261147adacd7b5e7b7dfc19d27511d6c66e60783ae5358a342efca",
|
| 40 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/candidates/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H15-L15x3-hrmCycle-Ref6-lowRealFinal-AdamATan2-EMA-pseudoBranches1.yaml": "e94a1b85805414a54de6d18ac143ebc2b273873938867aac73450f75f256c98d",
|
| 41 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/candidates/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H15-L15x3-hrmCycle-Ref6-lowRealFinal-AdamATan2-EMA-pseudoBranches2.yaml": "4f8153d360feab1ca2feb74caee3cad608864886f9fecc8b3b6f7df75a4622f4",
|
| 42 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/candidates/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H15-L15x3-hrmCycle-Ref6-lowRealFinal-AdamATan2-EMA-pseudoBranches3.yaml": "150f6683f843e4242ca7993ff8eca1451922aa8555cdc7e54c0581cb8f4d5182",
|
| 43 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/candidates/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H18-L15x3-hrmCycle-Ref3-lowRealFinal-AdamATan2-EMA-pseudoBranches1.yaml": "86f23ada4f759ed5f00a69c03683ad920d16307153a139757acd15169fbeb7cd",
|
| 44 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/candidates/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H18-L15x3-hrmCycle-Ref3-lowRealFinal-AdamATan2-EMA-pseudoBranches2.yaml": "2c7cfaba5b4e852c4f5defc8adc41028efc34970c99b1cc51214a46ce0b4b5f5",
|
| 45 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/candidates/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W1440-H18-L15x3-hrmCycle-Ref3-lowRealFinal-AdamATan2-EMA-pseudoBranches3.yaml": "b56ca440928a3328f4a498ecea233a290b6c377ed42dd99584a31b0d834cd8fb",
|
| 46 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/candidates/README.md": "0652caf008599a2bec944680bdd7c7e75f4020bc79ffccb69f0ceb30aa63d85d",
|
| 47 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/HRM-H2-L3-W768-H11-L8-BP8-causal-AdamATan2-EMA-seed0-1epoch.yaml": "273da800eee8183924fcb5db1822ad6004d3dce6281e9cc899ec81e2f9818dcd",
|
| 48 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/Lookahead-K1-LA2-A2-R0-D0-P0-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-Ref3-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "a2e318a403b0728bf943c8d42d79c76dee5639bed3eb2602e537db9bccc0b749",
|
| 49 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/Lookahead-K1-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-Ref3-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "749b1423e98c00baa8ca87a23acfc612311a232bbe0cfd5ff6ea803609988ff3",
|
| 50 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/Lookahead-K1-LA2-A2-R0-D0-P0p2-F0-seed0-W768-heads12-H8-L8x3-hrmCycle-Ref3-lowRealFinal-KV3_20_27-AdamATan2-EMA-1epoch.yaml": "d84de20f281e012d778e79e083ac1058c0e23baa0c840482acec34979a7665bc",
|
| 51 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/Lookahead-K1-LA2-A2-R0-D0p2-P0-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-Ref3-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "30fa807d983491ead62e296b4172bf1044689ed4f90073fc3185c13fa995cb2d",
|
| 52 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-Ref3-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "ed991e37d67f5f3ece1a502f86a5540a9de7cce54c2be8b505bea8972955f190",
|
| 53 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/README.md": "588d8f0b3954cb7f8d18ff234f82de0dc95d60918f0a155335acb2a841fdeaac",
|
| 54 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/cycle_depth/HRM-H3-L3-W768-H11-L8-BP12-causal-AdamATan2-EMA-seed0-1epoch.yaml": "a5517fb2812272cdd66617357765285551fdf106cd1e7f2d8470d37741c8bd78",
|
| 55 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/cycle_depth/HRM-H6-L3-W768-H11-L8-BP24-causal-AdamATan2-EMA-seed0-1epoch.yaml": "f29dfea6af46d1bc20222c19468ece25f92f70def6368a68fcd0b881d70b97c5",
|
| 56 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/cycle_depth/README.md": "65e281265a23a2532526b30aeaafcb90ffc7b13146325a6972070c22dbe66817",
|
| 57 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/double_compute/HRM-H4-L3-W768-H11-L8-BP16-causal-AdamATan2-EMA-seed0-1epoch.yaml": "64ff9bbdca6453389497087895869552ec50f61dae37556235e7507d6294e626",
|
| 58 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/double_compute/Lookahead-K1-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "3de6eae57f9418fbe7ab0bac3b0dc45a9a4e43527d7a55d9b1108294f11a5ad1",
|
| 59 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/double_compute/Lookahead-K1-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "bfce05b726ef39bda75d7426a85947e45c85acdc98dec5f58bd59f1c8f08775a",
|
| 60 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/double_compute/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch.yaml": "bbfab16d9670ac1d20ed07994425ab1fa1d6600576bc1e4259ee9be4e825ffad",
|
| 61 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/double_compute/README.md": "0a289855c0e8d06f294fa9de53dff9f57e1ce7431d2c9354f3c7f8b2a2ff9950",
|
| 62 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/r1/Lookahead-K1-LA2-A2-R1-D0-P0-F0p2-seed0-W760-heads10-H8-L8x3-hrmCycle-Ref3-lowRealFinal-KV3_20_27-AdamATan2-EMA-1epoch.yaml": "087fc824781a6a1d8c2ea256b24601522a2a3cee84c4030fda380b3ca684c2dd",
|
| 63 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/r1/Lookahead-K1-LA2-A2-R1-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-Ref3-lowRealFinal-KV3_20_27-AdamATan2-EMA-1epoch.yaml": "3246f631c80987e5ed2ec6a3d51b43c34535e507055f15e3a752176acfc4201e",
|
| 64 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/r1/README.md": "d015bfcbecbe04fd8c2fb8a125ef6867ac8c4a54e76f117dff6f5d85b524ce15",
|
| 65 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/r2/Lookahead-K1-LA2-A3-R2-D0-P0-F1-seed0-W768-heads12-H11-L8x3-hrmCycle-HC2-Tau2-N1-firstResponse-sharedBackbone-AdamATan2-EMA-1epoch.yaml": "2640ad3da6fcfa4da20bf76039719c3c60d58179e8983bc52b10535a8ebfaf3b",
|
| 66 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/r2/README.md": "83062014a46750b4791dead66ae99e1888e851f8b70d554a18baf0b6cefc2fc5",
|
| 67 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/refiner_depth/Lookahead-K1-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "ca8b26ea1c04aa5347927d43826541813c1eb2c262035f62a5b5dd15a0e00744",
|
| 68 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/refiner_depth/Lookahead-K1-LA2-A2-R0-D0p2-P0p2-F0-seed0-W760-heads10-H7-L7x3-hrmCycle-Ref5-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-1epoch.yaml": "b0a91868755de46cf4ed8e077f3eba6772b7dd46793123db709b1f2157369bf9",
|
| 69 |
+
"configs/causal_hrm_text/official_recipe/hrm_cycle/small/refiner_depth/README.md": "084164d47c5ec01161dd0d66ada5302c6e2194b016ca7a3f4f69633f8f673bc6",
|
| 70 |
+
"configs/causal_hrm_text/official_recipe/hrm_fbw/HRM-FBW-K3-H2-L3-H18-L15-causal-AdamATan2-EMA-seed0-SrcL1-L3-H1-L4-L6-H2.yaml": "756ef58acdb572a7064e3b77485406a00a40e6bd56059c010342b530da3a5f4f",
|
| 71 |
+
"configs/causal_hrm_text/official_recipe/hrm_fbw/HRM-FBW-K3-H2-L3-H18-L15-causal-AdamATan2-EMA-seed0-fromHRM-epoch3-SrcL1-L3-H1-L4-L6-H2.yaml": "b74b5ef3e5125bb971113b72d9a30129d8828309d731d1c5e6a7c62eb36b54f5",
|
| 72 |
+
"configs/causal_hrm_text/official_recipe/hrm_fbw/HRM-FBW-K3-H2-L3-H18-L15-causal-AdamATan2-EMA-seed0-fromHRM-epoch3.yaml": "b4cae4eb314b23f4f9baa5dec9b8d4eb11524c216e5da98335dfaef93bce2df4",
|
| 73 |
+
"configs/causal_hrm_text/official_recipe/hrm_fbw/HRM-FBW-K3-H2-L3-H18-L15-causal-AdamATan2-EMA-seed0-progressiveK-75-90.yaml": "df7c72acb6fd18b38d943aaed825480c5c21e5156731626e7d83cee0475785f1",
|
| 74 |
+
"configs/causal_hrm_text/official_recipe/hrm_fbw/HRM-FBW-K3-H2-L3-H18-L15-causal-AdamATan2-EMA-seed0.yaml": "a7d0297c78c1314a5b044a47c202b78536d469f05dd921ae236648f3638f5cd4",
|
| 75 |
+
"configs/causal_hrm_text/official_recipe/hrm_fbw/README.md": "6ece462cd4f44613126af32dd5273251083fbea402524d6c58f4868336e4fac4",
|
| 76 |
+
"configs/causal_hrm_text/official_recipe/hrm_full_bp/HRM-H2-L3-H18-L15-BP8-causal-AdamATan2-EMA-seed0.yaml": "160f3df51b5a6ccc645e96236894c93f37bd5449b25280be9977f7ea74b8deeb",
|
| 77 |
+
"configs/causal_hrm_text/official_recipe/hrm_full_bp/README.md": "734dbe39f67b0341b91ceda75090bfc5dd1ab2f262f40cea3b904ed70d749974",
|
| 78 |
+
"configs/evaluation/hrm_reference_v1.json": "602d9f37bef631c05d1a4a415cc5535a364b42df3eb6ae30a44d77b14ede7715",
|
| 79 |
+
"configs/fullbandwidth/README.md": "c7ff94c432aeb5caa5d7260374aa3099581b95394e8c8653e2307182bf397f86",
|
| 80 |
+
"configs/huginn/476M_7b.env": "ff98a658d92cd406b6edd4fa7d78f5dfd9f6b19d6804dbd40af9b66b9ba8363a",
|
| 81 |
+
"configs/huginn/Original-Huginn-0125.yaml": "2ace28d739e3dd978ab5f125f798aad9f6e964ae2ca0edf9840180e55851ea0e",
|
| 82 |
+
"configs/lookahead/476M_14b_K3_LA2.env": "c2854167570e9af2b01566c8385363d7cee5d4e37b054212a769a7dc97d7b0fa",
|
| 83 |
+
"configs/lookahead/476M_14b_K3_LA2_MB10.env": "c411538beed6b7589dda0e29362b6f0c64bfe193e947c32f9ef3a06ac5a72f47",
|
| 84 |
+
"configs/lookahead/476M_14b_K3_LA2_MB10_NORECOMPUTE.env": "3e41f5c085cc46b046c6013be4eebf9f44f6a3a04c92659b43025213b01606bf",
|
| 85 |
+
"configs/lookahead/476M_7b_K3_D0_P0_A2_R0_MEMORY.env": "1ab0c5d824acf347f44daa7f69da5dccbefbf9eeb3a3aff74924f9ac048c6cc3",
|
| 86 |
+
"configs/lookahead/476M_7b_K3_LA2.env": "da637667a121c3336190f08681eeeddc798723687279bc0df087daf0512ec2f1",
|
| 87 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0-P0-F0-seed42-refinerMemoryOnly-7b-compiled-a2bag-MB4.env": "07e95028d20298bd44f8d134ea5cee9d47ef0845e113e475ede303e35ca99f88",
|
| 88 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0-P0-F0-seed42-refinerMemoryOnly-7b.env": "77db16cf260947448464d2705c363fea8cef4e01c2da65dc558eeb79afd278c2",
|
| 89 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-refinerMemory-7b.env": "08b8a7b894495dffbe75f04bfe11045b43c5ab897db687640f9ca83ab88396af",
|
| 90 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-refinerMemoryOnly-7b-compiled-a2bag-MB2-auxAllPasses.env": "74d8829df7473a6516a8e3c77454dfa317e5b6732a44f60d95e37e5466af2b4e",
|
| 91 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-refinerMemoryOnly-7b-compiled-a2bag-MB2.env": "e79b81a20de8ef5bf7489caf67e7e61e63540e6d1ab9a8615ad096d8e99c09ec",
|
| 92 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-refinerMemoryOnly-7b-compiled-a2bag-MB4-auxAllPasses.env": "e38d2a3c7eb2369abca9f29d9fb51d97c01be436fe319a329fe4304a7226fc8f",
|
| 93 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-refinerMemoryOnly-7b-compiled-a2bag-MB4.env": "99ce2bd607c230d6fc6b61a427f990855f0c089bf236163eb7722f9eb4c30256",
|
| 94 |
+
"configs/lookahead/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-refinerMemoryOnly-7b.env": "b4cb96d810799641fea03e03bd544efe3d3ce44290919bc2070f5082325d4cf9",
|
| 95 |
+
"configs/lookahead/README.md": "677b9c76547bca8cac83dbd678831b150b7e1e811a5220a2f8aa89ca612c8b15",
|
| 96 |
+
"configs/lookahead/REFINER_MEMORY.md": "a5acc46c38705f3649019cff8ab9090cf492bdcd10fda539060ac0633794a94f",
|
| 97 |
+
"configs/lookahead/ablations/D0-A0-R0.env": "11f564edbcd7174def648a316acfa626bfb7b865eccc4d302b3e28a3540ddc49",
|
| 98 |
+
"configs/lookahead/ablations/D0-A0-R1.env": "67649480c5114ec93a6804ef65db81682e1e038436952b20ff24429899cbce1c",
|
| 99 |
+
"configs/lookahead/ablations/D0-A1-R0.env": "acedda67b5427523cabdbe33fa89612d307417985304133963a5c32ec0010ccf",
|
| 100 |
+
"configs/lookahead/ablations/D0-A1-R1.env": "b292ec0d0e34fb6f59bb9490e091906081ab5783c5120df3932530ce11a34f45",
|
| 101 |
+
"configs/lookahead/ablations/D0-A2-R0.env": "5fa99960fc5ed3f0163c4bb6ab1974ea62270f1db074a7f7793b50ff0637e384",
|
| 102 |
+
"configs/lookahead/ablations/D0-A2-R1.env": "462222782f87efa565da30e051362337478aa563e5210f1da7657bf9bc72a206",
|
| 103 |
+
"configs/lookahead/ablations/D1-A0-R0.env": "073c2dd201ee938cac8a2d292f7fd21a11101c25259a5b02039531058d60dd2f",
|
| 104 |
+
"configs/lookahead/ablations/D1-A0-R1.env": "68aeeeee676d3b2341319d8750c6f4c0b76727db565b562e89e1d223c5fe4380",
|
| 105 |
+
"configs/lookahead/ablations/D1-A1-R0.env": "98d6d580d93fe378d60b9390055446ec13e8c1077988ccf25c99c85c137d9c92",
|
| 106 |
+
"configs/lookahead/ablations/D1-A1-R1.env": "6f25e2327e0aad0dedfb44ba2248b47772af1c8afaed0f14fec502f5182578fc",
|
| 107 |
+
"configs/lookahead/ablations/D1-A2-R0.env": "a8548ec79fcd2382440d86a92bc4d2e7874e49c316f645cb487bcc7bb62851e4",
|
| 108 |
+
"configs/lookahead/ablations/D1-A2-R1.env": "9a782326b8dec0ad23d772ae9c6c77fa0ba328c34f0caa749bd19305bb00c30e",
|
| 109 |
+
"configs/lookahead/ablations/common.env": "509a7f31514ec0aa8b1954bf06da3af37ce15d7c29b78f24658f19763f708b7a",
|
| 110 |
+
"configs/lookahead/cpu_smoke.yaml": "47d67ca8299b2aa9c82dab3bff5558df271be3b8b0be0a38d838d28167bc9375",
|
| 111 |
+
"configs/lookahead/cpu_vanilla_smoke.yaml": "c0bbac9f37b0d938a5e20971b8d2a7c202252895caa2099a38d032cc33501812",
|
| 112 |
+
"configs/loop/476M_14b_FBW0_K3_L2_ADAPTER_SHARE_KV.env": "d6fd71c4ff02e9fd349f94c20224f87a30201e2cb90c07b293a00ca89e282fc3",
|
| 113 |
+
"configs/loop/476M_14b_FBW0_K3_L2_ADAPTER_SHARE_KV_MB10.env": "c7d9eb250df3e171eea91cef825bf2a8827d9efefe6d25300517dc4205d9e28c",
|
| 114 |
+
"configs/loop/476M_7b.env": "d462059d110e9d923660abc486811e25dd402d6489715ec5a490ec0b784529fb",
|
| 115 |
+
"configs/loop/476M_7b_FBW0_K3_L2.env": "0338ff980fe5ab6ae093d1335a808f4998baa589be5141b7262ec92534b54a0f",
|
| 116 |
+
"configs/loop/476M_7b_FBW0_K3_L2_ADAPTER.env": "ad1f5f6afb79e42a6cb80efe89da63f54ba5249982171907aa3bba619e34605c",
|
| 117 |
+
"configs/loop/476M_7b_FBW0_K3_L2_ADAPTER_IFM.env": "a5b0afb57b1d34b2c1cee722ff79fc6eeb4b7b246493f29fb0414397be23e7bd",
|
| 118 |
+
"configs/loop/476M_7b_FBW0_K3_L2_ADAPTER_IFM_SHARE_KV.env": "8573d72be7041dca69dea2d3d6cdaf7fc8e8ebf6b48f28dd0860ddaf806e4b72",
|
| 119 |
+
"configs/loop/476M_7b_FBW0_K3_L2_ADAPTER_SHARE_KV.env": "aae7032e8ed65417a5d673afa2e46a1941574342b39359b6473dbdb13b573ed6",
|
| 120 |
+
"configs/loop/476M_7b_L2_ADAPTER_IFM.env": "047976485fab887985c74af3b599189dccdf5caed056ce9ff6ad9c42b235eb8c",
|
| 121 |
+
"configs/loop/476M_7b_L6_ADAPTER_IFM.env": "7ac3f9618f4d7573586e8d58b2c50f33dea43b53ba156272e0e171764d2cebad",
|
| 122 |
+
"configs/loop/ADAPTER_CONTROL.md": "59b3c18cd11a3c7e9d6185ddffa58cbda0b3ed7d165ad26cc1c437f2fe418652",
|
| 123 |
+
"configs/loop/IFM_INPUT_WRITE.md": "ab201adfa6303859c0ea061fb34ea3396cca2b3c6c345bdf2cf16578d6c83613",
|
| 124 |
+
"configs/loop/KV_SHARING.md": "878b777bfc0c85e5e0c6244852e215fc0f1a335e3e87824ca984ce488a2204c6",
|
| 125 |
+
"configs/loop/Loop-K3-L2-KVFresh-Adapter3-NormOn-IFMOn-seed42-7b-refinerMemory.env": "c9abee3f4a4f27497c8d174d53e6ac110fe114cb367c841bf6598de0209a55ee",
|
| 126 |
+
"configs/loop/Loop-K3-L2-KVFresh-Adapter3-NormOn-IFMOn-seed42-7b-refinerMemoryOnly.env": "5330d12d8e3be2503232680771434780084f771a0d56ecf1a7a398bd39f3c1bc",
|
| 127 |
+
"configs/loop/NO_FBW_BASELINES.md": "130499e6c3bfe0daa6556b2ec9d47a95139f700a87e6329202847c688a7811ee",
|
| 128 |
+
"configs/loop/README.md": "a92cca650e18eb7ed8d2845c3707b0c75b6a48f91b038641836acbfcb29f409e",
|
| 129 |
+
"configs/loop_budget/Huginn-K1-P5-R10-C5-T11-seed42.yaml": "686293173d71e8c40eaa120f03f4b58816617060a4e9819c35b861dbaa540a91",
|
| 130 |
+
"configs/loop_budget/Huginn-K3-P5-R10-C5-T3-seed42-FFN5568.yaml": "830b1ed25eea7df5a6727157cd7bcd261285c9cc561d3fd482cf8163423ab730",
|
| 131 |
+
"configs/loop_budget/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly-auxAllPasses.yaml": "f0f69894d17419423adf5fe39aa15300ef8de53719c38f56da98441f689f8a52",
|
| 132 |
+
"configs/loop_budget/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly.yaml": "f6d10a0a56af7cc0df36ebe50d33ef8119d68f6b8a7d3eea88dae5438a650795",
|
| 133 |
+
"configs/loop_budget/Ouro-K1-B20-T6-Stage1-seed42.yaml": "99229cab2f4f2e5bb63189066eee1130085f69c066e114daff1ac7021a1ac786",
|
| 134 |
+
"configs/loop_budget/Ouro-K3-B20-T2-Stage1-seed42.yaml": "d7d1148c34dcd8a50b8484d3b4ccfc36a54bb6b56b1c93daa2d3bfb4aa2ad58a",
|
| 135 |
+
"configs/loop_budget/README.md": "19b7937c9cdfe4bde98177a22e79cd7ad70018ae068d06d56cbc01085fb91f41",
|
| 136 |
+
"configs/loop_budget/Vanilla-B20-seed42.yaml": "b3022a1ecb52c99c7449dd60731b5535015b8c36745a001a05c247a480d6b50a",
|
| 137 |
+
"configs/loop_budget/common.yaml": "6787aa6f4c2384e78b98e18cf1179b12521e4149ea6ab7f3a3817a698f45e333",
|
| 138 |
+
"configs/loop_budget/horizons/20tpp/FBW-K6-B20-seed42-20tpp-linear.yaml": "e5cbb5e6fa373b38d9b3e7473e10129fe3fe96b4927091528e2b45f9ab18f40a",
|
| 139 |
+
"configs/loop_budget/horizons/20tpp/Huginn-K1-P5-R10-C5-T11-seed42-20tpp-linear.yaml": "e9f9f635f520dcd441902c359400104270207fd5aaee7cdc08dba5b6c11cbe0a",
|
| 140 |
+
"configs/loop_budget/horizons/20tpp/Huginn-K3-P5-R10-C5-T3-seed42-FFN5568-20tpp-linear.yaml": "cbd82f7533f3028f12c9618176f51a95f175ec430a5fb69ef8970856eca227f8",
|
| 141 |
+
"configs/loop_budget/horizons/20tpp/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly-20tpp-linear.yaml": "15f9bb24c94e0ae77abcc976310f12f98d15b98634ebcfab540fba433f20defc",
|
| 142 |
+
"configs/loop_budget/horizons/20tpp/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly-auxAllPasses-20tpp-linear.yaml": "a8d73a3983e3157f15bf338d49f332204978aca2b25a9e67d8db296ea9a99ee2",
|
| 143 |
+
"configs/loop_budget/horizons/20tpp/Ouro-K1-B20-T6-Stage1-seed42-20tpp-linear.yaml": "f2bc50d78775bd3b7d9d67d5f8997bc471e0b41824fa70f7225a53391b739d22",
|
| 144 |
+
"configs/loop_budget/horizons/20tpp/Ouro-K3-B20-T2-Stage1-seed42-20tpp-linear.yaml": "d3073319a6a6bfcd6491f171c4c5f67406de7abed61bbb9229f2a8261df7012c",
|
| 145 |
+
"configs/loop_budget/horizons/20tpp/Vanilla-B20-seed42-20tpp-linear.yaml": "72ee7275fe5cc595c754cae2626adbc12d30f14697a4045105e3010107ea0550",
|
| 146 |
+
"configs/loop_budget/horizons/40tpp/FBW-K6-B20-seed42-40tpp-linear.yaml": "b406f2faae2aac6882b792019a0edb2aacbbae6694068d0a553893c9bcd71e74",
|
| 147 |
+
"configs/loop_budget/horizons/40tpp/Huginn-K1-P5-R10-C5-T11-seed42-40tpp-linear.yaml": "41b75eb152bff4a93575a21e4ef76fd9e4dab456b7c2161fa5cdb72b8710fa41",
|
| 148 |
+
"configs/loop_budget/horizons/40tpp/Huginn-K3-P5-R10-C5-T3-seed42-FFN5568-40tpp-linear.yaml": "16ce948fbd00a794e2b7d1cfdcf8848e8501d998da3a641c5a82337c77b626f0",
|
| 149 |
+
"configs/loop_budget/horizons/40tpp/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly-40tpp-linear.yaml": "9e16cced4f791fcd6b9126d0bcb5bc49570b5ff7862ab7acdcf12303dba30805",
|
| 150 |
+
"configs/loop_budget/horizons/40tpp/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly-auxAllPasses-40tpp-linear.yaml": "d293852c834153f38b7d206a499a277d8c0ecb49e241ae67041db79e8b05b7ea",
|
| 151 |
+
"configs/loop_budget/horizons/40tpp/Ouro-K1-B20-T6-Stage1-seed42-40tpp-linear.yaml": "3225a452dd583cfcf4ff61ec1a96c34ac073cc16d09ebe1d0b21230eedfc8689",
|
| 152 |
+
"configs/loop_budget/horizons/40tpp/Ouro-K3-B20-T2-Stage1-seed42-40tpp-linear.yaml": "5ebfa4c39cd204f025c9616f811c885c2840d9e384de36c9cd48636bfd9e99c0",
|
| 153 |
+
"configs/loop_budget/horizons/40tpp/Vanilla-B20-seed42-40tpp-linear.yaml": "112f6db0c5319d5910f39c4e68e2a6a6b3d1f55ed2af4887fee597fe809de576",
|
| 154 |
+
"configs/loop_budget/horizons/80tpp/FBW-K6-B20-seed42-80tpp-linear.yaml": "c6f397f4fa938c59bfb5c2635af658d6ad0057d13c3909f918c3a15cc121943c",
|
| 155 |
+
"configs/loop_budget/horizons/80tpp/Huginn-K1-P5-R10-C5-T11-seed42-80tpp-linear.yaml": "ef7102edbd1cbe0ac299e2a94027b4ac3f81a64bcfa87820a332c55d11287739",
|
| 156 |
+
"configs/loop_budget/horizons/80tpp/Huginn-K3-P5-R10-C5-T3-seed42-FFN5568-80tpp-linear.yaml": "d0ce0a1a6df8fb44f08d1d8b76fdbc534643ee0e2ce67077b003379b745d7486",
|
| 157 |
+
"configs/loop_budget/horizons/80tpp/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly-80tpp-linear.yaml": "f8bae0f2dc879a5a03f10a691ac3f85f58a3ce347ec437c2a04d2fca5cbe516f",
|
| 158 |
+
"configs/loop_budget/horizons/80tpp/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN5120-refinerMemoryOnly-auxAllPasses-80tpp-linear.yaml": "467c034ff78bc4e6f036a2263d3c055f8aaad29c18ba9fa9ac9621dc5215ff07",
|
| 159 |
+
"configs/loop_budget/horizons/80tpp/Ouro-K1-B20-T6-Stage1-seed42-80tpp-linear.yaml": "81eab705e09efaa0ce1f9767305a2642ddb5b6a1c48f9deb9b84b80739307e4f",
|
| 160 |
+
"configs/loop_budget/horizons/80tpp/Ouro-K3-B20-T2-Stage1-seed42-80tpp-linear.yaml": "37413752861da41a8eae4e4ff465f1b96a8b5f8b6d5e7a2f3e349277d5406f17",
|
| 161 |
+
"configs/loop_budget/horizons/80tpp/Vanilla-B20-seed42-80tpp-linear.yaml": "e6cd543cfa0b0fe02b8a97c3f35b01349a905e48f36d3a9af0acc847388e0586",
|
| 162 |
+
"configs/loop_budget/horizons/README.md": "9cd28ee1272ae94e69b8cda7502c46347915e68d23a36444039f4eb5148b4872",
|
| 163 |
+
"configs/loop_budget/lr_bs_search/PROGRESS.md": "98dd8870a1ac3de4fa21d4ca6a6cd7b566372ed85af10234ca396c1bbf5613c5",
|
| 164 |
+
"configs/loop_budget/lr_bs_search/README.md": "6083a2ba176893fc81d4f0cc10a2ed6ed99b6bc3b7bc17254487848007502a0a",
|
| 165 |
+
"configs/loop_budget/lr_bs_search/RECOVERY_R2.md": "2d51326aba384391e90fea2488c4d7d5d8b45000e3dcaa39e35e9e6c32848dc7",
|
| 166 |
+
"configs/loop_budget/lr_bs_search/arms/FBW-K6-B20-seed42-50m-1b.yaml": "3f01673018886ff2cb01b7e88a3d58bbebc712665ba68bf532f01021b4703648",
|
| 167 |
+
"configs/loop_budget/lr_bs_search/arms/Huginn-K1-P5-R10-C5-T11-seed42-50m-1b.yaml": "1c9cd43b8c9fad49308935d7aa83b53af15d499a2b803611bda4f0145899cd7d",
|
| 168 |
+
"configs/loop_budget/lr_bs_search/arms/Huginn-K3-P5-R10-C5-T3-seed42-50m-1b.yaml": "14b3568bad86b6c4fca76f21cd2556bf4746352c03fab1f620fd8a4e41d21e54",
|
| 169 |
+
"configs/loop_budget/lr_bs_search/arms/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN1536-refinerMemoryOnly-50m-1b.yaml": "e5019be2d863ce88cbacdff053d997b2287de533fc19dd0032a30b3d932e13d0",
|
| 170 |
+
"configs/loop_budget/lr_bs_search/arms/Lookahead-K3-LA2-A2-R0-D0p2-P0p2-F0-seed42-B18-Ref3-FFN1536-refinerMemoryOnly-auxAllPasses-50m-1b.yaml": "c4cbd050505e1401ec4a9646222dac1e46a1e1c177b72a1fe6047ca747a34f49",
|
| 171 |
+
"configs/loop_budget/lr_bs_search/arms/Ouro-K1-B20-T6-Stage1-seed42-50m-1b.yaml": "4fa9ddb0cb617746f0738b73094a1f4950a02b61ab50cca910cc3668f3290f02",
|
| 172 |
+
"configs/loop_budget/lr_bs_search/arms/Ouro-K3-B20-T2-Stage1-seed42-50m-1b.yaml": "0f0270bcd0e11dc15abf7d7a2e63229ff7cae9ab0de72b3f65a235d104c5028c",
|
| 173 |
+
"configs/loop_budget/lr_bs_search/arms/Vanilla-B20-seed42-50m-1b.yaml": "54618ea6e5d730d07e9b850b6d9bc75efa929047fce4edcf25dda28cfa610109",
|
| 174 |
+
"configs/loop_budget/lr_bs_search/common.yaml": "904c38957fe2924e7c9c615305e925c87d0327aba235d41df8713ea0a4754ca0",
|
| 175 |
+
"configs/loop_budget/lr_bs_search/recovery.yaml": "3be7734fa4de6230139761366cbf7f157167f6ccb70e954fdd002d9f2dc01cfd",
|
| 176 |
+
"configs/loop_budget/lr_bs_search/search.yaml": "2d5799cea3cc8c038b63223a6076d1337a42170c2e2fdc418f1207ac007b18da",
|
| 177 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS131072-MatrixLR0p5.yaml": "d624a3102ee21c13cc531773bd864d1236c20d79ca1d6b8246fbe35d0d5084e8",
|
| 178 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS131072-MatrixLR1.yaml": "31ea4077e09600634d063bbb5e71b2ff279e21b28b053c0940dd966a216ed67a",
|
| 179 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS131072-MatrixLR2.yaml": "f85bdafc5d7d98453b0af32118bc92ff9dabf1737f9c5f0dee7dfc79ec87ef9e",
|
| 180 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS32768-MatrixLR0p5.yaml": "383623a91ef4abf5ed700fe61cc710f66d47a06d1410053fe31696d79e7ddc89",
|
| 181 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS32768-MatrixLR1.yaml": "9f7d2c87782cdbbbd07562e5cd0d072ce5b33189ff5894e76a6c438f8a1e1a10",
|
| 182 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS32768-MatrixLR2.yaml": "bb6cfe93065c567853793aeac50d5628e11880bffaeee478a6275f28f8a817d5",
|
| 183 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS65536-MatrixLR0p5.yaml": "9582a8dce8171e4c96858664ae25a5bbf68d2fe88a3656431aa0c7fde621d74f",
|
| 184 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS65536-MatrixLR1.yaml": "3e2e820cdd50fa601ad360a3e6f7f6a683102b784e05ca2b7c454ad73d60b90e",
|
| 185 |
+
"configs/loop_budget/lr_bs_search/trials/Vanilla-B20-seed42-50m-1b-BS65536-MatrixLR2.yaml": "0bf7d4bdd0d25a4f923919a8d999e570cb2a2605d3839e1cfed6941d3474fabe",
|
| 186 |
+
"configs/loop_budget/scripts/README.md": "d96252d9befa3de46533756abcc7ad463db65253b39c759b844ea2eb4b991263",
|
| 187 |
+
"configs/loop_budget/scripts/lr_recovery.py": "16b5a6e8ee9584ceb93dc02336d62589ce0acd30ea27c874f5ba377b17e2b407",
|
| 188 |
+
"configs/loop_budget/scripts/lr_sweep.py": "f7face633e88c7df27319eede3c1d7a298a62121f5a2c597f90a296307f3ac92",
|
| 189 |
+
"configs/loop_budget/scripts/prepare_sweep_data.py": "ff4949fe9930747c54d0983d05996e1e5baa4394918e5820af89c6c097bcdbf4",
|
| 190 |
+
"configs/loop_budget/scripts/run_lr_recovery.sh": "5cd46245b748f0b3262be0ba56bd33c63178e2a913fe600dd854d2c42f420725",
|
| 191 |
+
"configs/loop_budget/scripts/run_lr_sweep.sh": "16e11331e256975b13f9b17cac29444686601e915574eb123a0d2a07785392ea",
|
| 192 |
+
"configs/ouro/Original-Ouro-1p4B-Stage1a.yaml": "f3f3ac5b80fba1cbe87a597214c17bb83f16a1a9fabdbd9436295ce8b052ad48",
|
| 193 |
+
"configs/vanilla/476M_7b.env": "4b6078d83d75ee005d08daba368834373550cb1a50c8108435a4a210e1dd6580",
|
| 194 |
+
"configs/vanilla/476M_7b_FBW0_K2K3_50_50.env": "7610ddb814c4b98b22ec08bad94bf18bb8367fde6648b32e4d817c34a8126e28",
|
| 195 |
+
"configs/vanilla/476M_7b_FBW0_K3.env": "3de80a7f024da0e64da9706d6475ed28796cc7632f967e6bcefe1abf8eebb10f",
|
| 196 |
+
"configs/vanilla/476M_7b_FBW0_K3_23L.env": "4b062d2053cea91f2e9f20a83d6fb450bcb059fbc50bc3e2f73934ae3b1a00da",
|
| 197 |
+
"configs/vanilla/476M_7b_FBW0_K3_L2L11L20.env": "637c50c798de3f8796a5112318af3c2ce772a32fd22aa9e2ea11d38b11dccca9",
|
| 198 |
+
"configs/vanilla/476M_7b_mtp.env": "273e3bbcf3e42f841c98b057b89129b7c00b07095df5c38b6818298533e41db2",
|
| 199 |
+
"configs/vanilla/476M_7b_mtp_FBW0_K2K3_50_50.env": "3f6d852dba48a32d81b6b52a7ee12a90b9c4ef1a29a6dad67025c73ef6380ea8",
|
| 200 |
+
"configs/vanilla/476M_7b_mtp_FBW0_K3.env": "909ee411f3d7395c97db22eb350c9f92a2ba515bc0ce8402103f3fedb78e9547",
|
| 201 |
+
"configs/vanilla/FBW_DEPTH_COMPARISON.md": "696b4acd24d207208eb38ddae1761d46d4170747db2304276cab4b6a6f0eeb4d",
|
| 202 |
+
"configs/vanilla/FBW_MULTILAYER_STUDY.md": "58203cb1a1a45f4cf750c4605ef267836a3359c665bd46a57a54461c90456f74",
|
| 203 |
+
"configs/vanilla/FBW_STUDY.md": "5430f5a4bc3883856edb22d4a198264dedb4477666eba4ce748e2708841d2307",
|
| 204 |
+
"configs/vanilla/MTP_FBW_STUDY.md": "83cafc81e90d1e6ec0071a4fb34db1395c6c0ba0bbc4be744c49f8c31c7d6b7a",
|
| 205 |
+
"configs/vanilla/README.md": "fc622c1361e81a176ca3b031a6b0a865a18ab28752011f60c59df29c381d0b2e",
|
| 206 |
+
"scripts/hrm_official_worker.sh": "f842ed079c2262cae6314e2870d377fb66d39884c54c8c6bd234bf3604704112",
|
| 207 |
+
"src/full_bandwidth_loop_tf/__init__.py": "7d7213a2ad8c47a51bd619e58185bd6a77bbb54dd3d42c098031e6bb980d7994",
|
| 208 |
+
"src/full_bandwidth_loop_tf/adapter.py": "f7979fb2c58d185aea94af08c042798c1e18c690331363bde05d13f6e595533b",
|
| 209 |
+
"src/full_bandwidth_loop_tf/adapter_policy.py": "7944aedde11852f2821e77d325c6963f3537ac18eb7d7a6e9deee38867ef7193",
|
| 210 |
+
"src/full_bandwidth_loop_tf/hf/__init__.py": "f9f4f6d9cfaa80a0443983c70e46cec8b9361ea69a75667717eb8df1c201cc74",
|
| 211 |
+
"src/full_bandwidth_loop_tf/hf/backbone.py": "0eafcedbf30552ea910628b5e845ecb73d9dc2daca99d73e38748d873e3b42dc",
|
| 212 |
+
"src/full_bandwidth_loop_tf/hf/config.py": "f82ca8335df23b35b22ba63d0bc42437c97e3cb0d960943b6608e3c1cf4f6827",
|
| 213 |
+
"src/full_bandwidth_loop_tf/hf/model.py": "21219220c8c4313fb8a6290bf5d492631f135dba679af407151b3f6ca924e683",
|
| 214 |
+
"src/full_bandwidth_loop_tf/hf/plugin.py": "2b15947ce20c8be1676551ddcce6be9bf65d0c3f200064b06fe0ae7460a7f4b4",
|
| 215 |
+
"src/full_bandwidth_loop_tf/hf/task.py": "400c0869028798ed1badc4b6d7acebe917fc0f490b22eefbc0f150e10ce5305c",
|
| 216 |
+
"src/full_bandwidth_loop_tf/hf/train.py": "5767c6d879066068860b173a0b212ed82928ab42e5bbbd8c93bdd3cc032533fe",
|
| 217 |
+
"src/full_bandwidth_loop_tf/input_write.py": "a7974576081355741c3188d38508f78da95db913347cd5a60396f98932d4b05b",
|
| 218 |
+
"src/full_bandwidth_loop_tf/input_write_policy.py": "51c68dbc21d7a2fdd604669d97a464207cb5ad0200dfab3b0ef9ed80edfa723a",
|
| 219 |
+
"src/full_bandwidth_loop_tf/kv_reuse.py": "7ac602c587d10b83768e067ec7c0cc358d924028ca76abff7209f411853074c2",
|
| 220 |
+
"src/full_bandwidth_loop_tf/megatron/__init__.py": "82327efe57c404760d9b8cf9c083e85c9e526fc1c53e782d156adaa66b9b3f60",
|
| 221 |
+
"src/full_bandwidth_loop_tf/megatron/adapter_decoder.py": "aea500799671712ebcd6683cb1fb0fccd648791d2ca103e9caa5050da787a542",
|
| 222 |
+
"src/full_bandwidth_loop_tf/megatron/attention.py": "6835e2149abec4abb848b52406466a12d654cab2bebe19c129a679f64c82f91c",
|
| 223 |
+
"src/full_bandwidth_loop_tf/megatron/builder.py": "3d8391ce8c65c721acd1e6fdefc07bdb94d888ffea042eb88b6fe10fdb702854",
|
| 224 |
+
"src/full_bandwidth_loop_tf/megatron/config.py": "1c5801721a7a75de8bc65d138e0c0dbeee86d4113bb20938ab664e2672ba4839",
|
| 225 |
+
"src/full_bandwidth_loop_tf/megatron/decoder.py": "bea4c2ff62dfebcbf08a156176cfe65f8d31c7df1fa24ea3c4991421d1e99a14",
|
| 226 |
+
"src/full_bandwidth_loop_tf/megatron/model.py": "dee02d33cd9c8d389f71af71bf7e53d73ab5a886107335ca8b335bb57f0cb01e",
|
| 227 |
+
"src/full_bandwidth_loop_tf/megatron/plugin.py": "ca8ee90c1e547ba7935c46fd2c2f2cd436940151c4915a4480eaa60243033728",
|
| 228 |
+
"src/full_bandwidth_loop_tf/megatron/task.py": "c28c28caae4cc9e098d6a7a600acac970d6fc10928063cb4a1be5264106dfa92",
|
| 229 |
+
"src/full_bandwidth_loop_tf/megatron/train.py": "4ffb371d68045560a0c48297fdb686592573e44dac826fbd4605639cd317f4b0",
|
| 230 |
+
"src/full_bandwidth_loop_tf/policy.py": "a80e630571be5468a0ef0e1788cd7d6520e8dc158bcdfe6576bf47fe6917e2d2",
|
| 231 |
+
"src/full_bandwidth_tf/__init__.py": "73932d7b3105e983f86321784ff547181848cfde537630ab60e93b96c5c80729",
|
| 232 |
+
"src/full_bandwidth_tf/hf/__init__.py": "6d9ae0e0360910223451d706e71477e258a4bea6ebaaa99287c77abc4b8a0130",
|
| 233 |
+
"src/full_bandwidth_tf/hf/backbone.py": "da1062836060d9ea1580366ba4ecea04e409da1b5f23d328047040d44a5154f0",
|
| 234 |
+
"src/full_bandwidth_tf/hf/config.py": "dabcc7e08f9d7026515d67393e0803f59c843598c59f790ec5934d6154cb8c27",
|
| 235 |
+
"src/full_bandwidth_tf/hf/data.py": "9c9c3f53eae324c0c2acac63898b155f555f0c65191349c906d7530ebd0ab0dc",
|
| 236 |
+
"src/full_bandwidth_tf/hf/diagnostics.py": "40dff542e9d801872b1e1bd2dca35bf5f39aa38dd1ece67f44c1b81aaf72cd92",
|
| 237 |
+
"src/full_bandwidth_tf/hf/generate.py": "75fabd6e56c0d19de30a0743623277e7ef1f1c242fb069b6ec5acefe19d030dd",
|
| 238 |
+
"src/full_bandwidth_tf/hf/inference.py": "9424da65353b2a55a2ecf179bcbf3e3d7c5fee735b6951cf8b356f4621dd7894",
|
| 239 |
+
"src/full_bandwidth_tf/hf/model.py": "fa9ce4e58510b4a373fc8ed11ba96a87abfef8e8a4ef7fc9c495aedd46744aaa",
|
| 240 |
+
"src/full_bandwidth_tf/hf/plugin.py": "5f9cf723a71a2b1db9a3a03f395be964a4e175f2a60d493872c136fb2987035b",
|
| 241 |
+
"src/full_bandwidth_tf/hf/task.py": "523d09295a271969a1b54369e6344be02850052c83f0135b3becb929b8d65139",
|
| 242 |
+
"src/full_bandwidth_tf/hf/train.py": "3c83db8e6a91088fe00e781a69d218f2d0c4ae1317a8ecd86cbb33045d3002e2",
|
| 243 |
+
"src/full_bandwidth_tf/inference.py": "6e100bc0ee8792bfe85b0042e25e24a04d433b133377da301ab678f6b6f26a5d",
|
| 244 |
+
"src/full_bandwidth_tf/launch.py": "eb60929434506c1fafcb62a450fd68daf14ea578aac40a3b2e7ca88589ae5646",
|
| 245 |
+
"src/full_bandwidth_tf/megatron/__init__.py": "0ecf759960f5e1d412b3939b54394734df3ee54662fb73cbf621b9c773a8b3ef",
|
| 246 |
+
"src/full_bandwidth_tf/megatron/builder.py": "770af80866b19ac6bf785c61658b874ebd14b5ebecbcc3699410ac8ca252b290",
|
| 247 |
+
"src/full_bandwidth_tf/megatron/config.py": "1279e4ce26146a9879bce34eaeec2e7f53392a79cebf949d2942b70f9a1d2441",
|
| 248 |
+
"src/full_bandwidth_tf/megatron/metrics.py": "bb7fb343d92ed6cd95a5bc091dfad26dfe94b9f4fb9c9dd22d05353f2ec62938",
|
| 249 |
+
"src/full_bandwidth_tf/megatron/model.py": "da485c9cb837535132804c6e3a40d209f26fa9cdbd335450dd518d306f23e8e8",
|
| 250 |
+
"src/full_bandwidth_tf/megatron/plugin.py": "ddc601d23017d25423f67b309aadaba1272c0fc11a4fc53c755f754f4317cb0d",
|
| 251 |
+
"src/full_bandwidth_tf/megatron/task.py": "99bc52999dbd72285252b855bd61709c3388dd2a3dcf47e52c643cc8c052ce67",
|
| 252 |
+
"src/full_bandwidth_tf/megatron/train.py": "7da747ddd915c361b04efd18dbdd6afc99243dc866be0fbb71b596460f49e4d5",
|
| 253 |
+
"src/full_bandwidth_tf/mtp.py": "77c6217551615384289fe676187414bff35a229f89032cb734a32d649a8c5cce",
|
| 254 |
+
"src/full_bandwidth_tf/policy.py": "ccef928f79fd85b5b270f9d4c1c7a32b30151fb2e3751432109bd0406aa52d5a",
|
| 255 |
+
"src/full_bandwidth_tf/validation.py": "935f19865a4e5b34962d21ef3c00d3c3614acd094bf2fcba579d0779eb8765ad",
|
| 256 |
+
"src/full_bandwidth_tf.egg-info/PKG-INFO": "121e6cc95ed808b9805f999423eacd8dc8dcc538641d8de7da3db8a347758466",
|
| 257 |
+
"src/full_bandwidth_tf.egg-info/SOURCES.txt": "99ed1edf6d05339ccdc8b23e6bbfd9d58cb536ef4080a4f1616d4a58ea5eba08",
|
| 258 |
+
"src/full_bandwidth_tf.egg-info/dependency_links.txt": "01ba4719c80b6fe911b091a7c05124b64eeece964e09c058ef8f9805daca546b",
|
| 259 |
+
"src/full_bandwidth_tf.egg-info/entry_points.txt": "44998f7c1bc5c962ad977370cdc1eace976e66b1c13c2b93e50bd274fef45eeb",
|
| 260 |
+
"src/full_bandwidth_tf.egg-info/requires.txt": "60a7f93a0205d64fb030f984c29d1135fabb40925402ccd5d92be370fea4a473",
|
| 261 |
+
"src/full_bandwidth_tf.egg-info/top_level.txt": "0212ad8598b39436853cd1881f4e815d69e4892925d6af74359f3e223eaae5aa",
|
| 262 |
+
"src/hrm_text_tf/README.md": "ca71a8bd52ab9f13cba7e5af3a264ad99a5ca7d9726a1467c12ec43e4ade507d",
|
| 263 |
+
"src/hrm_text_tf/__init__.py": "620200ffcbd80720ba089f4ab4ab7e72ec4f6b173a39d1861afdf48402021fc3",
|
| 264 |
+
"src/hrm_text_tf/check.py": "053fc64435041813f7a33bf8908b0246d0879ba440a4fe91340b808af88af86f",
|
| 265 |
+
"src/hrm_text_tf/hf/__init__.py": "70a45fbb9b72166003150ff69545dc8a143ca408fe26e6ed216b2381cb9359ac",
|
| 266 |
+
"src/hrm_text_tf/hf/comparison.py": "abca6729c41be1fd23ff9f6d91612329d25f5db144f88ab61082647a0c2b43af",
|
| 267 |
+
"src/hrm_text_tf/hf/config.py": "ffd0453b21509a02b546163d490b05060f843d0f08383bf09ce7f5d45b7418bc",
|
| 268 |
+
"src/hrm_text_tf/hf/data.py": "7190165dc31c52f5e248b6d61b3be84b952c6ae45e758e37df1f6cbe1a3bdb9b",
|
| 269 |
+
"src/hrm_text_tf/hf/lookahead.py": "d1985cb6a4be48ff51a2198eca757281bc5d0d94a19c18c57f7a42f8209d1aed",
|
| 270 |
+
"src/hrm_text_tf/hf/lookahead_backbone.py": "e5ba82d4c2bdf9dd3c976ab5409df73d35c40ee32eb16d1041225bec09659e65",
|
| 271 |
+
"src/hrm_text_tf/hf/model.py": "46119c39416c98ab59517ac8fe0468b83c31e173e8f3493d42ed5cbdd93c7572",
|
| 272 |
+
"src/hrm_text_tf/hf/plugin.py": "db2a59aa0d0b0e1c6644d44c6cb1fc65123124b9683f4fb56b42a713c0fc85ed",
|
| 273 |
+
"src/hrm_text_tf/hf/task.py": "32d0c615d455afbe51753c59c15452ec61370e97b560ae2d93c3fdce08961107",
|
| 274 |
+
"src/hrm_text_tf/hf/train.py": "6625025b4d741b08641f71252d17871b8ce95c3b365d8f34e904542e0bca8b4e",
|
| 275 |
+
"src/hrm_text_tf/official/__init__.py": "2e261493323f21b1f3727bf36fcf3ddb7ae51b90564772ea660f454c9f3c5dd4",
|
| 276 |
+
"src/hrm_text_tf/official/attention.py": "9a23cf86481be95601b188d4f3aa013d34c15c8e0caf341b9a5b2e7dce3b15b5",
|
| 277 |
+
"src/hrm_text_tf/official/block_execution.py": "7d1688fbf68a7610b3330c8c000d7a2529fbd5fd1dd7bd5abaca0cf4813a847d",
|
| 278 |
+
"src/hrm_text_tf/official/config.py": "c4e9d1f67ef476243138fbf3fb34efe0b88576967e8ddcbcee8f794e238a144f",
|
| 279 |
+
"src/hrm_text_tf/official/continuation.py": "740408e1476a0c9f7f3c6af685196b3d7fb8ed1f222177dfdf3bca7c542df6bc",
|
| 280 |
+
"src/hrm_text_tf/official/evaluation/README.md": "101e4f52df818847be90d4a40d605d6c688394a25754e55983df4628114c74a5",
|
| 281 |
+
"src/hrm_text_tf/official/evaluation/__init__.py": "e513c05e06297db1b7e6588e0dfb97627f71891f3bb7c87cd160652ddc96a702",
|
| 282 |
+
"src/hrm_text_tf/official/evaluation/aggregate.py": "1624fa9aeac027e4a09706dd731805d112978ddcb82acff7843c4f0b0d934c4d",
|
| 283 |
+
"src/hrm_text_tf/official/evaluation/attention.py": "bc12635a76e03be1deed5fb08ff2f31858f8a797a7999558f397591919b1e4a9",
|
| 284 |
+
"src/hrm_text_tf/official/evaluation/boolq.py": "0646f823feb2ec6fb3728fb603ef71ab862617766a04c5b9d2f96df97f896414",
|
| 285 |
+
"src/hrm_text_tf/official/evaluation/checkpoint.py": "3a2389afd25396270defa4827ce8d82ef6970ee995b132c9e59fb16a88f54794",
|
| 286 |
+
"src/hrm_text_tf/official/evaluation/code_results.py": "66f8958a4ba12d390b9581561190e5a20c8bf2b523b2ce8f4982c97ac0661001",
|
| 287 |
+
"src/hrm_text_tf/official/evaluation/cs_algorithms.py": "a9ab93cabd2101ebfd25dfac9e865e4569a9396a983db961d897527d20c9bd6c",
|
| 288 |
+
"src/hrm_text_tf/official/evaluation/extended.py": "73ec525acd626b37c1d4ab2a54f7df60af0087385c555e15e703bf4207241efc",
|
| 289 |
+
"src/hrm_text_tf/official/evaluation/fbw_generation.py": "fe8c3550262d88b7eee6e231c96d5d2328e9a92483b66db3cf7d4eeba6b66599",
|
| 290 |
+
"src/hrm_text_tf/official/evaluation/generation.py": "dfff17dbb931c937ce7256b5cdb25f732df7951fde1328ec7a8b3ca011393dd8",
|
| 291 |
+
"src/hrm_text_tf/official/evaluation/lookahead_generation.py": "01b67a41d1262b5211f65e0e69793a58b2eb44d66213569a9a8a24055dd69924",
|
| 292 |
+
"src/hrm_text_tf/official/evaluation/main.py": "cd5ede756c31198dd6a559a987a5834fdb679da5917d297d0a6fb04cb0e9f63f",
|
| 293 |
+
"src/hrm_text_tf/official/evaluation/prepare_reference.py": "ea708aee653da45d18339a632070004b1ffc5cb3cc93b7690878c585a6a60cb9",
|
| 294 |
+
"src/hrm_text_tf/official/evaluation/queue.py": "d0b3f070270c3fcf480f63f4b3c186b4e0c4658a939318617ea260c63fd4dd63",
|
| 295 |
+
"src/hrm_text_tf/official/evaluation/reshard.py": "e0954de5f4bc0afd7b61d31d5effb690950774d4949eb69962ee79ee9df3267c",
|
| 296 |
+
"src/hrm_text_tf/official/evaluation/resume.py": "75aa87768fc32013b55852996d51a189cf5e534d96826f7f5ea8615c86fa00cc",
|
| 297 |
+
"src/hrm_text_tf/official/evaluation/suite.py": "ba1fd5c636969ea89aeec56b5f3ef907216a158adbf437e702c5c5d2c7e3b9b6",
|
| 298 |
+
"src/hrm_text_tf/official/evaluation/webq.py": "b8d0daad1acb0992aff38c6b2698b4b0469bde5e4603b8293ba379c23d3706a9",
|
| 299 |
+
"src/hrm_text_tf/official/gradient_clipping.py": "876de8e63768667668aafd3bcb5fb685b4fb2c29c7775779a5cb9fef1eb0cd19",
|
| 300 |
+
"src/hrm_text_tf/official/hrm_feedback.py": "19fda3e0e9821c6f77b45d935adb6a36a0a010767ade899686002ee8c9fab1a1",
|
| 301 |
+
"src/hrm_text_tf/official/lookahead.py": "dd9140600923a76e95848fdbddd0758848265a519474135cb55d8a3c1562b417",
|
| 302 |
+
"src/hrm_text_tf/official/pass_schedule.py": "56f680210d53c842191495f3cd7fb4e1d0efc59c5475c0457e091b6607d31ef3",
|
| 303 |
+
"src/hrm_text_tf/official/r2.py": "cb0938dfd7fe701eba56fba7f73eb912677ae8b8e58a074483b7a175e45cc089",
|
| 304 |
+
"src/hrm_text_tf/official/redistribution.py": "0eb1e430f38ae0ba2fe5c5ff8468a7dde33dc006559ec5cbe257c9687787ec78",
|
| 305 |
+
"src/hrm_text_tf/official/reference_options.py": "25c7b69ec7325fe580c05ab9881d7e1063b10b3fb3a3000ce7df2246561b14b1",
|
| 306 |
+
"src/hrm_text_tf/official/reference_validation.py": "a685f737796837fdb2495e1fb0bd3f144051d630179b39f8cb3be72358cbd115",
|
| 307 |
+
"src/hrm_text_tf/official/resume.py": "b4196413d1af8a8e7330ede547b8c3f475c6bcd7dae10ac4c9abee37b7b019e5",
|
| 308 |
+
"src/hrm_text_tf/official/runtime_cache.py": "babc34ecd141b97490cb691376ddb637660e822d1cf936b8139a8d4f8544e031",
|
| 309 |
+
"src/hrm_text_tf/official/train.py": "075f29629b2168f4d821c8bc71fb5c417585ee8f062438c1b13cca1dc730c3e6",
|
| 310 |
+
"src/hrm_text_tf/official/training_budget.py": "a557c63845de1dcda427918a5737b5e59bedfd755384d044293dee3ca99ca38a",
|
| 311 |
+
"src/hrm_text_tf/official/virtual_replicas.py": "eb3161208e2862e5a6ef0429ac9f17bb7488fec96baaa17a11074aeecf0d91a3",
|
| 312 |
+
"src/huginn_tf/__init__.py": "39d5e260fe7267a636c4e674a00972c727b74cd961e80cc9ae3af76fadf5d213",
|
| 313 |
+
"src/huginn_tf/hf/__init__.py": "012202a3f5f7ad94fb75141e9e7acd572e736d35037396403b937071197a8f58",
|
| 314 |
+
"src/huginn_tf/hf/config.py": "988917571eda3428ebbcb3b396dbca1040f78350bb6eb20d844d1a5573949307",
|
| 315 |
+
"src/huginn_tf/hf/model.py": "e55e1b0b4755c517175818f74629a9b04c047547e95b09cf3408c49c31a337bd",
|
| 316 |
+
"src/huginn_tf/hf/plugin.py": "f13fb4625549b47af55be1f1ef3200cfdebeed16be3d5fb0784d8eab14d41dcc",
|
| 317 |
+
"src/huginn_tf/hf/task.py": "ca04d330852ba8c0c190be72d95c8802a65701fed9848ea3af8d608f7c85e6dc",
|
| 318 |
+
"src/huginn_tf/hf/train.py": "a3f8facc58f3d13e835fa36829d8b8ee72123ac098604f1ce4be615e8743e8f6",
|
| 319 |
+
"src/huginn_tf/megatron/__init__.py": "01a90198bdfa78ebf53ee3cf69078c7c6c11ccfc9dfdbbb840348c0453d884e3",
|
| 320 |
+
"src/huginn_tf/megatron/builder.py": "a30735ff0ffe0df2340d8b0a351dd82792cd2ee0dda5488954f0e4dec3d954e2",
|
| 321 |
+
"src/huginn_tf/megatron/config.py": "82731a0f18fc14825672c085d8febb395bb4196a60177453d852f8297ecd3f2f",
|
| 322 |
+
"src/huginn_tf/megatron/model.py": "208d1ce0da3753171efe2cdd2aee01e882bf067c1135c4ac4ef7a1cce8adeee3",
|
| 323 |
+
"src/huginn_tf/megatron/plugin.py": "7c1b7e54758ac416d5f4af1cc0f79c7ff98c292d1f5721c55246486ae33e5093",
|
| 324 |
+
"src/huginn_tf/megatron/task.py": "43f8ed00f51e7bac2176c5a35060b4d12ad8a11f5b46dd3aaca690a599963a01",
|
| 325 |
+
"src/huginn_tf/megatron/train.py": "d5636e052befdd38e4f971c8a7f38b7178cac783c6295e2ca787f8fa742fb5dd",
|
| 326 |
+
"src/huginn_tf/policy.py": "05393214dbb2f9abdf0c1b317d5254d7888e9de52411f62c345cbe90fbb987ca",
|
| 327 |
+
"src/huginn_tf/validation.py": "7dd329bbf0bef8dc287f2d337cb66cc479ba90cd7e138f5fcc25ed1881f10671",
|
| 328 |
+
"src/lookahead_tf/__init__.py": "58894edf8fb480516d6667e25e88fe198f830b8de2cf9e6b71bd926137779a28",
|
| 329 |
+
"src/lookahead_tf/compilation.py": "f3412e21d1eaf5136396ab5363fc65cde243156a49d3d4f1db80c1d4dd43fa39",
|
| 330 |
+
"src/lookahead_tf/hf/__init__.py": "75016b1cc8b2fae544c7fdbbfa5c7d0511217a63261223f0920fd6acf3be3bea",
|
| 331 |
+
"src/lookahead_tf/hf/attention.py": "f06e224885a7211cf057d7c3968b16b2a0ebcc3fca1822e6afbd958022ec8d78",
|
| 332 |
+
"src/lookahead_tf/hf/attention_reference.py": "79c0db25b0667cc0f8128d224009ce85cc9021f7d0b36acb1e55654fdc631975",
|
| 333 |
+
"src/lookahead_tf/hf/backbone.py": "3da59b4c4a1bd2df0b79ec4d1c70be81076c4aa9f38eda1d3381f2f7b3fbeecc",
|
| 334 |
+
"src/lookahead_tf/hf/config.py": "67bf0756a3eee7e02086b37587e14bebf022651e49faabd6d4e592405d9637b3",
|
| 335 |
+
"src/lookahead_tf/hf/decoding.py": "4b6cc64f578c411cce44e754252afeb591f6fb861befc8e0519d016833e447a6",
|
| 336 |
+
"src/lookahead_tf/hf/diagnostics.py": "1942539a5947dcbe3c1ca15d7cd8558c0538453a0444c9bd1fa9646c1ab1cd7a",
|
| 337 |
+
"src/lookahead_tf/hf/embedding_mix.py": "4de03db21ff4f3a3bfea1269b6eba91277a53903b1c9dfa3d55d83241a540101",
|
| 338 |
+
"src/lookahead_tf/hf/inference.py": "f604e68bde0697a18dd8b8a2894fd712a8c7adaac54d36f47a25eff6662496b6",
|
| 339 |
+
"src/lookahead_tf/hf/k1_inference.py": "a6eb3c653587d5c9007938480c5f2edca8827fa391283cb38d5eb8a3b1fce06a",
|
| 340 |
+
"src/lookahead_tf/hf/k1_static_inference.py": "0dc312f16efb84965b680531b2c8dac1f2e948ce4ca3fd4687a629cf562a0481",
|
| 341 |
+
"src/lookahead_tf/hf/losses.py": "f2e60c334be2b96a0967e3faf4bce28b049ec806bcccadcb4d7aa458758f41be",
|
| 342 |
+
"src/lookahead_tf/hf/model.py": "8968e75959c5ff206fcdd35ba19e6429650b1440d922ac819eff0df80386adac",
|
| 343 |
+
"src/lookahead_tf/hf/paper_backbone.py": "6147f1b50a5e5d30844973537a455784ed60385cf7cfdba54029043a93a13ce5",
|
| 344 |
+
"src/lookahead_tf/hf/plugin.py": "85576c47247a5267af087bb863a3b3f5b9c8743eb3dc6d8138000edd2f0040f5",
|
| 345 |
+
"src/lookahead_tf/hf/refiner.py": "d7d36e1538c57a21c7985cf54203bd7aa39111908b5e871ca40b40c93c0a3c13",
|
| 346 |
+
"src/lookahead_tf/hf/state.py": "e45abaeb8968ea869fce0d2b9595213f4e441db34c0eaa7e69291ae026e58afb",
|
| 347 |
+
"src/lookahead_tf/hf/task.py": "5ea98496e38a614504a31d41aa33279fd875b8377133fbabe3d75713cf10ccb3",
|
| 348 |
+
"src/lookahead_tf/hf/train.py": "5d63387f67c4a8bdeb3b641e2736c5fd1825cf66ba2c3069fe79320f39bd5670",
|
| 349 |
+
"src/lookahead_tf/hf/training.py": "698fb321dc1145c4c25898426f3c0014456540e74c8956568fead98f64b6bb9a",
|
| 350 |
+
"src/lookahead_tf/initialization.py": "82743d0f658019149f9c3119cf314c3a2f4dc45d110d566595f63d52c00e859d",
|
| 351 |
+
"src/lookahead_tf/megatron/__init__.py": "83ccc5a2293c077268c8bf855672b1020caf8d1396c338f8570af629a4456a1e",
|
| 352 |
+
"src/lookahead_tf/megatron/builder.py": "a605d8b146f6a73f225a87bf981294a5ee6d89ac11b7ab00221b8e68b17c22ee",
|
| 353 |
+
"src/lookahead_tf/megatron/checkpoint.py": "5ba116561ef87094b7c208d884d1f98af030d8df47d11965231528d5816ab2c4",
|
| 354 |
+
"src/lookahead_tf/megatron/config.py": "6d264fd773d664c1f613b86ec8f6faa4c7e7b71c97bf7df7ba0042b094e3fd08",
|
| 355 |
+
"src/lookahead_tf/megatron/model.py": "01a3202748c1558cb75d2d14ba484a1fe811bd41e5e3881a859c0287ba75c6ea",
|
| 356 |
+
"src/lookahead_tf/megatron/plugin.py": "c2a6e985600641a77e173ecbfceb561c9c23aa98e3b2ed7509fb75f59907fedf",
|
| 357 |
+
"src/lookahead_tf/megatron/task.py": "2e6d660e454ef8153e3c73cb64a656e45232fe60ffaed4fb39c577c6f429a412",
|
| 358 |
+
"src/lookahead_tf/megatron/train.py": "b5ecabeac41b2791dbf7a88530610e2540f87a7d851a1bb4671e4b9ff07f75be",
|
| 359 |
+
"src/lookahead_tf/policy.py": "37b8bbe6b46d6b911ee03539636b510a5fe26e434855a2220890370791b4c510",
|
| 360 |
+
"src/loop_transformer_backbones/LoopTFDR/README.md": "742136ba9b9c6d8e020a6721469c833b95f3b33551957333ba9bcc83864ecb25",
|
| 361 |
+
"src/loop_transformer_backbones/LoopTFDR/__init__.py": "1d8fcf503f0d78a23e448155572aa57201c357b38e7cf1b563bd064b64ce197c",
|
| 362 |
+
"src/loop_transformer_backbones/LoopTFDR/layers.py": "6ff5451060c7ab8113b004021d3e59d40ad41556e0dafabf564dca8717d315f5",
|
| 363 |
+
"src/loop_transformer_backbones/LoopTFDR/megatron/__init__.py": "410bb1030d04a7a8bc692b422e5e52466699d96123581e05c8dcbb4d7814320b",
|
| 364 |
+
"src/loop_transformer_backbones/LoopTFDR/megatron/layer_spec.py": "691ab807a84b9023cab4a28a296695f6fb6cbffcc3e1d3b124760d7c0f0f3833",
|
| 365 |
+
"src/loop_transformer_backbones/LoopTFDR/profile.py": "794fccdb5767e1b8325b466893fe087a2638e206d2b6c692af5e21b82c34653b",
|
| 366 |
+
"src/loop_transformer_backbones/LoopTFDR/torch/__init__.py": "dda23e33011a0dc06211a9a093c92a341feb82e870819b50ad719487967f4626",
|
| 367 |
+
"src/loop_transformer_backbones/LoopTFDR/torch/block.py": "3e178c8c11aadc72cdbf215f8bd8717b93a0157147b236989480376317e183d2",
|
| 368 |
+
"src/loop_transformer_backbones/__init__.py": "8517ac48322900e373518ef0c557229f02a9558a71003820f5c9b21ef3b5d587",
|
| 369 |
+
"src/loop_transformer_backbones/attention.py": "775e6571e467fdffdde3815d02c647edf675611905d5de9539fdd661b4131492",
|
| 370 |
+
"src/loop_transformer_backbones/hrm_text/__init__.py": "38a673898d1b96a9137eb6f369752e734575a63a95d722a1511d5f9855c8d765",
|
| 371 |
+
"src/loop_transformer_backbones/hrm_text/block.py": "e62e1100483496caf9678001248b68ca52ca4521788db0ec64e6e4cd4e449bef",
|
| 372 |
+
"src/loop_transformer_backbones/kv_cache.py": "6657158add8b3fce3ef5b5c7f1ba6700ba0ba4642de36101a6ef2ff2e982d994",
|
| 373 |
+
"src/loop_transformer_backbones/packing.py": "fceb4a6017017a428945bca322d8a684c46e439578c5eebeeed497c7510dbdda",
|
| 374 |
+
"src/loop_transformer_backbones/paper2604_21106/__init__.py": "3458d444e0795105ca9c40a0136be6616eed25d0511de4ea4bd73efb2204d57c",
|
| 375 |
+
"src/loop_transformer_backbones/paper2604_21106/initialization.py": "530597dccacf6669057b75f52ee7b8654b7722e9a32295d27443bc56f205ed9d",
|
| 376 |
+
"src/loop_transformer_backbones/paper2604_21106/layers.py": "882edb527b62820e794ccb5d8617d76934e41a0725f25fdb0463d8edf0c008c0",
|
| 377 |
+
"src/loop_transformer_backbones/paper2604_21106/megatron/__init__.py": "3458d444e0795105ca9c40a0136be6616eed25d0511de4ea4bd73efb2204d57c",
|
| 378 |
+
"src/loop_transformer_backbones/paper2604_21106/megatron/initialization.py": "3eacb55bc58ce6263f40e9b5ff5aaa4c47fa69b045c2ffe0bc00788386f5ec0f",
|
| 379 |
+
"src/loop_transformer_backbones/paper2604_21106/megatron/layer_spec.py": "a1bb0f4b424664acd33069c4efc1fe67287cf2624c7425691df85aef348fd75b",
|
| 380 |
+
"src/loop_transformer_backbones/paper2604_21106/profile.py": "6c36cd38f3a8d3445a2d1a3d53da013f771e3a3474cca5e4abe10a6f5df64f08",
|
| 381 |
+
"src/loop_transformer_backbones/paper2604_21106/torch/__init__.py": "3458d444e0795105ca9c40a0136be6616eed25d0511de4ea4bd73efb2204d57c",
|
| 382 |
+
"src/loop_transformer_backbones/paper2604_21106/torch/backbone.py": "dcace7df22bf1498ee7436ee3a6fc08f74d790b63150142da2572708f6b13602",
|
| 383 |
+
"src/loop_transformer_backbones/paper2604_21106/torch/block.py": "fa6667637901dd3581d486c5d02cdf42b33f56817e17dba26cd165841e41643b",
|
| 384 |
+
"src/loop_transformer_backbones/paper2604_21106/validation.py": "4cdb33b4e1d359bf4d588903249f26c1eb21f647108f93966f43d30121314aba",
|
| 385 |
+
"src/loop_transformer_backbones/paper2608_08888/README.md": "8f39f91e5d2c9797fdf95da79b74f2c433b276a94a3ee3c62534595063f7ab78",
|
| 386 |
+
"src/loop_transformer_backbones/paper2608_08888/__init__.py": "74583d5f333a24f0914acd3fe8ed8c8549d026f0cd73f9ed469e14749f91ead2",
|
| 387 |
+
"src/loop_transformer_backbones/paper2608_08888/backbone.py": "cbcb042c08de4ba7bf9128cda9cc295e39f9f63ff1e416c40f621052c97c02e9",
|
| 388 |
+
"src/loop_transformer_backbones/shared/__init__.py": "928c302cd7510b6cc04eeb0c955b19f31f5fdb5136d77748206673a2aae185d5",
|
| 389 |
+
"src/loop_transformer_backbones/shared/attention.py": "d4b15037eb5011abb9777e8c7fa274fb2f83baf0fcdc916543878d573309ae7a",
|
| 390 |
+
"src/loop_transformer_backbones/shared/rope.py": "1515180bcc9ba71bd31b2d514bd8402d0a1a42931cf76b27a923051a35ad1cb2",
|
| 391 |
+
"src/loop_transformer_backends/README.md": "10e4599c13cb8befbc576dae9b593c01022a5fe3256ba1a1475478722e1996a3",
|
| 392 |
+
"src/loop_transformer_backends/__init__.py": "1f97d268753cd796d5c663bf7a6e9e709f193efef88ecaa1fb463fd26ce8ab29",
|
| 393 |
+
"src/loop_transformer_backends/backbone_losses.py": "7ebc1618d504cd2370c1792e16fe404a0b68950c320bef250802142001e1f90e",
|
| 394 |
+
"src/loop_transformer_backends/benchmarks/__init__.py": "112d1ace8221c58c5798e5b40ad52c2fddf3f4c450016cac0fb8948b38750502",
|
| 395 |
+
"src/loop_transformer_backends/benchmarks/cli.py": "bb1bbde391ea814550dcdf6d69ee63b8daa5e1e9d824e0d08046e688ae70d4be",
|
| 396 |
+
"src/loop_transformer_backends/benchmarks/code_scoring.py": "20c635f8d4739b788d82ed427798b2163c94b7e58133fa70d24889207213dda7",
|
| 397 |
+
"src/loop_transformer_backends/benchmarks/contracts.py": "3a16b7684fbc78554c0c5f6ad83689da00dd8fb2e7d4f0322090d65555dfeac1",
|
| 398 |
+
"src/loop_transformer_backends/benchmarks/harness.py": "77f800da638dfb295c2bc47b064f75f3e304e6830f06e8cb2cc121de58a95bb2",
|
| 399 |
+
"src/loop_transformer_backends/benchmarks/ifm.py": "83cdfa5444f9efde75234ff0c0e55ccd76225d61bf213159845a3a7cf0ee72ad",
|
| 400 |
+
"src/loop_transformer_backends/benchmarks/ifm_data.py": "48fac62274cf48f3320f4fa2917c252d7b057db376646ec810c3de02d4319332",
|
| 401 |
+
"src/loop_transformer_backends/benchmarks/inference.py": "3a18955200b4d6ecf96022a3fc606bdd540bc7d627d4760f70025206fc9ced0e",
|
| 402 |
+
"src/loop_transformer_backends/benchmarks/models.py": "566eb3cdbea0efa480330c7dbc7c9061de5ab6a094074894f627bc936d7ee89f",
|
| 403 |
+
"src/loop_transformer_backends/benchmarks/native.py": "b9db22d4dd0db2f021886d3493a3094022664bc9b43b3f89221c01cf277dc631",
|
| 404 |
+
"src/loop_transformer_backends/benchmarks/paper2604/__init__.py": "bc7b4f54523aebc6bbb1493799eb8c044542187174c50adc41b07234b8e86558",
|
| 405 |
+
"src/loop_transformer_backends/benchmarks/paper2604/data.py": "f04ddc0f3d0923d2d4943cd612d16245e4b29ff4e9039abcedf283b2e07f3951",
|
| 406 |
+
"src/loop_transformer_backends/benchmarks/paper2604/models.py": "a30f13226af60138571be0fc8b6059485e91f7c6487f13da0830dc6218b49e67",
|
| 407 |
+
"src/loop_transformer_backends/benchmarks/paper2604/prefix_query.py": "c6280a79d5adf0eb69e6eb0c67602b7eb91c0b030b23a6ac7a1e553435e752ee",
|
| 408 |
+
"src/loop_transformer_backends/benchmarks/paper2604/runner.py": "c0afca509646e7d2ebaf23f076d1990ed91df6ae6787e00453a04eff224ca69b",
|
| 409 |
+
"src/loop_transformer_backends/benchmarks/paper2604/scoring.py": "18655cf51a3741f491862903c77c69bd4eaeceb474c6f9f13816e1f7a86fb522",
|
| 410 |
+
"src/loop_transformer_backends/benchmarks/paper2604/vendor/README.md": "454cbe8b8813b00922cfb121bc839813b3bbabc2b7d3dae09942e70eddc80f75",
|
| 411 |
+
"src/loop_transformer_backends/benchmarks/paper2604/vendor/__init__.py": "17ca169fe2554e86a286e9b02c93d5b500b2790e3c61a333bea576189dbd1791",
|
| 412 |
+
"src/loop_transformer_backends/benchmarks/paper2604/vendor/core_eval.py": "d2fd43fd18b0764a15c05110238ff92903cf15dc0bbb33222f6d83b059db2b65",
|
| 413 |
+
"src/loop_transformer_backends/benchmarks/paper2604/vendor/prepare_owned.py": "d855ab38b32f4510384dd14a7f4da433b8f96d7132811a08e436758ffe63b166",
|
| 414 |
+
"src/loop_transformer_backends/benchmarks/paper2604/vendor/prepare_saunshi.py": "63bf92ea02014c182861cf6b0ace09c99e06a57d83db46d9885ce8c3b6763ec9",
|
| 415 |
+
"src/loop_transformer_backends/benchmarks/receipts.py": "df963e734ccb94e4b1731bf9afef3f1f6c160ad4b82fa981e2b5f72eb57405ae",
|
| 416 |
+
"src/loop_transformer_backends/benchmarks/task_sources.py": "67a520c0ad121949eabbfa1d8c1702df3b47960e758279958cd6f64f20c4447e",
|
| 417 |
+
"src/loop_transformer_backends/causal_hrm_text/__init__.py": "66b9d23e61abd58ea554b8e96c6a333157f37051b20fb6ea8991d69fbc2bc8aa",
|
| 418 |
+
"src/loop_transformer_backends/causal_hrm_text/check.py": "c6e76a3613916cf442a8879fdbd9139b4ad68a6dc19c008e888ae039d3a942ea",
|
| 419 |
+
"src/loop_transformer_backends/causal_hrm_text/config.py": "17aaebd942ef355247fcffe340ef55b7d348ca8313d3dd58b26e7c3a63c79159",
|
| 420 |
+
"src/loop_transformer_backends/causal_hrm_text/data.py": "1ca1a7944d4baa663f8c7819b1c844b1e6aaf409a78067a48142a2e77834dbed",
|
| 421 |
+
"src/loop_transformer_backends/causal_hrm_text/model.py": "9aa24f4dfbdd53c76ebf3fd30f5dc020b75f0a08c772e7b6fcdad77ffcf330fc",
|
| 422 |
+
"src/loop_transformer_backends/causal_hrm_text/task.py": "ca3ce7017467da482705e7901cf35971958e28ebd7e815981fde5190e8cc96c7",
|
| 423 |
+
"src/loop_transformer_backends/compute_budget.py": "61ebe853d3cad27ecf1b5fa9c26dbde45993d80566899bd50664e59246312fd2",
|
| 424 |
+
"src/loop_transformer_backends/defaults.py": "52f9aa466cfc3d6dc5919d43a9b52f07ad9ff68d84388328a3debb57ad0bac18",
|
| 425 |
+
"src/loop_transformer_backends/evaluation_queue/__init__.py": "59585dc8311358fc00f265e56906d73ec9c20a9acce497a22d37fc66d5566c3b",
|
| 426 |
+
"src/loop_transformer_backends/evaluation_queue/campaign.py": "8f3592b936e80602e5f80c1658b58d9677042d48cab1ee209562271f0b14e527",
|
| 427 |
+
"src/loop_transformer_backends/evaluation_queue/catalog.py": "6448e63bc579adbecd121d6a3441d86d6c156c7d627ae9368fd5a5353ec60484",
|
| 428 |
+
"src/loop_transformer_backends/evaluation_queue/dispatcher.py": "aef771f76bf5f7fc694103de66eb8ce9616ec51daad4b4010d26a2863c947a55",
|
| 429 |
+
"src/loop_transformer_backends/evaluation_queue/harvest.py": "2c2cd8f1801c254f7ad308b5308a357c221756f8c9a1706fd127d4465d3dec30",
|
| 430 |
+
"src/loop_transformer_backends/evaluation_queue/naming.py": "85dad0a8f2b74db07d8d1b67696ff7fa677308dd95335a396cd257a081c0ee78",
|
| 431 |
+
"src/loop_transformer_backends/evaluation_queue/processes.py": "7a335e763bcb1409489ee569188e007548c288ef3c3fd352c14ef89e29f8c172",
|
| 432 |
+
"src/loop_transformer_backends/evaluation_queue/protocol.py": "d90ad08fa56a45842cb1eaf34ed6db14c48659dd98862d4e4c7b646b247fc3bf",
|
| 433 |
+
"src/loop_transformer_backends/evaluation_queue/shared.py": "49d2e14689d5d90d7dc2a839291e043b6fb4120c40f9e82970c4bc4b6755d293",
|
| 434 |
+
"src/loop_transformer_backends/evaluation_queue/split.py": "3ad95ea1be5bb8501fa2537ba995a8d8a0f9194a4c78d9ada23d3585d38e7691",
|
| 435 |
+
"src/loop_transformer_backends/evaluation_queue/worker.py": "ccc87849193ec7cbd4d80e0056bbd071f7c2d349dc9a8e81b428e4f0d9c59d0c",
|
| 436 |
+
"src/loop_transformer_backends/gate_diagnostics.py": "2697400d9eed000e5533c103155b038d3c90f98ffaf24cf1ee4084ccb9626a60",
|
| 437 |
+
"src/loop_transformer_backends/loop_budget.py": "d1a561ddab8b2881155a6849870af99526094b2fa463cf034b38414117c0bf74",
|
| 438 |
+
"src/loop_transformer_backends/loop_horizon.py": "4b33fff6f8a7db10d3cc14ba4cdf47dbee343d2a8f6d3d912bf948ddba427c6b",
|
| 439 |
+
"src/loop_transformer_backends/loop_search.py": "9495435ea0c6b36fb780a9e3f8c1c26f74bf580daf28f446ceddce67dd0e161d",
|
| 440 |
+
"src/loop_transformer_backends/megatron/__init__.py": "e34071777fb85eec3ff4c785ec243229259c9f74c1cbd0fc5281285653212838",
|
| 441 |
+
"src/loop_transformer_backends/megatron/compat.py": "b15db2b52c81221878ea3eb4f86d2829cb9d13da52a5bbdef26bdfe64af350e8",
|
| 442 |
+
"src/loop_transformer_backends/megatron/contracts.py": "05f4528ef5c07371caea28cfe563b8240f71a051d079e9d5ae11e041ef3447be",
|
| 443 |
+
"src/loop_transformer_backends/megatron/data/__init__.py": "95095cdda3d00fff34919f2058f5eaadcdc458944e43d7d21cc85adc6b75f2c8",
|
| 444 |
+
"src/loop_transformer_backends/megatron/data/batches.py": "0149909bfc246aceb268f092bf84e856bcedf55be6478f8a58395cbf50bd27e1",
|
| 445 |
+
"src/loop_transformer_backends/megatron/data/datasets.py": "117405f41d47870028852a8fbd8de4a452b18c357f61efdd8c4dd2f48a5ac6a0",
|
| 446 |
+
"src/loop_transformer_backends/megatron/data/online.py": "152ed6230fdaf4e5542fb3ff40045887d07c21ef47db820fa539d5dfc276f463",
|
| 447 |
+
"src/loop_transformer_backends/megatron/data/prefetch.py": "ad239e2a7760a7cc590afd446160fc1e808dc4cdb93d011b071427c13a571ada",
|
| 448 |
+
"src/loop_transformer_backends/megatron/data/raw_cache.py": "190f2b8d4291aadac92f9f23a9a19ba37806853c56a92bbf0cc85654057c3708",
|
| 449 |
+
"src/loop_transformer_backends/megatron/data/streaming.py": "86c472554205f1ce6317833fa43410a1ce5a7327599ec711ed67052821e90bea",
|
| 450 |
+
"src/loop_transformer_backends/megatron/data/validation_cache.py": "3b40979c919faf2b5e0e85f62db97ea4dc544431134f00874158cf7849c256ba",
|
| 451 |
+
"src/loop_transformer_backends/megatron/evaluation.py": "4fa012b45c8014ef067f2669231789178be63c6c0084c7d687203c3c66526f79",
|
| 452 |
+
"src/loop_transformer_backends/megatron/gate_diagnostics.py": "216693ee5b2c9baba709d3ab3b9d3f77345593b80840882c0af091432bdfe013",
|
| 453 |
+
"src/loop_transformer_backends/megatron/losses.py": "a258b7833d072537e3b2100b06709ec8c19855e2b71a82eb799190006cb6aeef",
|
| 454 |
+
"src/loop_transformer_backends/megatron/metrics.py": "174f60c0fa806472884f4ea3e4cfde08e9549acc04114fef3c84e25ebc9d7441",
|
| 455 |
+
"src/loop_transformer_backends/megatron/recipe/__init__.py": "c4e2d48cb2beca7b37ba30496be169f11092e80cdd65eb8ec6b318a76dd1f857",
|
| 456 |
+
"src/loop_transformer_backends/megatron/recipe/legacy_normuon_wsd.py": "ba41ca5350b032f1e44c62c9d254d9ce500bda1a15ccddc59dfa162431fbc9c5",
|
| 457 |
+
"src/loop_transformer_backends/megatron/recipe/paper2604_21106.py": "69df1d414019e2352021021b7926aea199b7c95a53dc075b10edeb2c2f098fc2",
|
| 458 |
+
"src/loop_transformer_backends/megatron/recipe/paper2608_08888.py": "6a706d05c2db4b7dae1fcf51408e2d466c5e58cc957d98512d65b2f3d107088f",
|
| 459 |
+
"src/loop_transformer_backends/megatron/recipe/shared_configuration.py": "1095f8ed7a40ed5f037f764fb5750223194792898f578e3e455b4f50bf7af907",
|
| 460 |
+
"src/loop_transformer_backends/megatron/recipe/shared_schedule.py": "5ad080fc1b52a66284d3eab3f15d709ce23c752c4528ae65b648cfa56bc8bf15",
|
| 461 |
+
"src/loop_transformer_backends/megatron/registry.py": "6c74d9e843b9d97c4ecdbdc678af2046aa8d1e5062190156864cf4c05ad472ac",
|
| 462 |
+
"src/loop_transformer_backends/megatron/runtime.py": "5f9565662c24893c0baacf5a4a335f795e79d29acc2156afcc46978145db38c1",
|
| 463 |
+
"src/loop_transformer_backends/megatron/validation.py": "b8965ab1459e023025447c5b56e3e1f3eaa85c430e3fd1845257e04d3e451284",
|
| 464 |
+
"src/loop_transformer_backends/megatron/wandb_hashing.py": "430208baf0399ae9d40b0211494802abb33a1f60845dad69e62a1310d3d09261",
|
| 465 |
+
"src/loop_transformer_backends/recipes.py": "7eb45d66610b5317fbe7d01831e7cc02529ad458727ab29e9fa6ba29f70e74a4",
|
| 466 |
+
"src/loop_transformer_backends/tokenizers/__init__.py": "c28a006de78276efa4110498c7c50b729469487670a80cd3da2b8895447a1ce6",
|
| 467 |
+
"src/loop_transformer_backends/tokenizers/paper2604_21106.py": "3a3df8031d26bd9032c3b508d9990168e1b15c2bf8a1c4703ccfdb6ae7cb5e73",
|
| 468 |
+
"src/loop_transformer_backends/torch_distributed/__init__.py": "c759ae93f0dc7bd282165417a46370c27ef61fcbd9ee96ba33843654ea67ff26",
|
| 469 |
+
"src/loop_transformer_backends/torch_distributed/checkpointing.py": "5497b7086eb329e2c52fd7238367960d42dd72a39d9ec67eb1f3dbb383de9060",
|
| 470 |
+
"src/loop_transformer_backends/torch_distributed/config.py": "658470133fb34e35ecbdccc5fd6f4cf93b03df20fbed34a65728290a1ae7d521",
|
| 471 |
+
"src/loop_transformer_backends/torch_distributed/contracts.py": "bf4de3792ea1664b61e7ca54d7cfa20476d542a07abe14c454a9e0b278de1264",
|
| 472 |
+
"src/loop_transformer_backends/torch_distributed/data.py": "0633000a1e98c0469a6503c18bc0ca0b022ed4672e4d1e1cf06d07582322bae9",
|
| 473 |
+
"src/loop_transformer_backends/torch_distributed/distributed.py": "ca634bf0f658fb8a9b635416c45dd75618a7d92bad9eed9dde02643c26f5615a",
|
| 474 |
+
"src/loop_transformer_backends/torch_distributed/muonh.py": "c5fc541a0b25ded840dc31aa2769c0b72122ef46c1019162d754c7240bff2d52",
|
| 475 |
+
"src/loop_transformer_backends/torch_distributed/optim.py": "4cccf133f01c65ba9821e25f5bef8530eb8aca96c29846e4985c8f09088b59ef",
|
| 476 |
+
"src/loop_transformer_backends/torch_distributed/reductions.py": "42db3763bcb69352a9a4c592c2da0135300230692f741a4e50a8468e78cafc19",
|
| 477 |
+
"src/loop_transformer_backends/torch_distributed/registry.py": "04297374a0f6747ca235193642ca2bbd90df1375a3c119be4496f6efe2832198",
|
| 478 |
+
"src/loop_transformer_backends/torch_distributed/runtime.py": "60793a222ddca43cf6d68b5f4d43ed040651a2a4b717f851d7cf715b6185e32a",
|
| 479 |
+
"src/loop_transformer_backends/torch_distributed/schedules.py": "cfcf6e60e5ca17eb105e351535ff87ae85440fe0d78beba6dfc03799b9345379",
|
| 480 |
+
"src/ouro_tf/__init__.py": "a73fdac27f075f98269937ecaba86e6c5e3b4bf36764f1350a6b5efddd25081a",
|
| 481 |
+
"src/ouro_tf/hf/__init__.py": "b1c80c4f72cca2e9870ff2cfb717923ab2e29365dd4e91f3b484a8e9eebbf432",
|
| 482 |
+
"src/ouro_tf/hf/backbone.py": "4f7cbaa3bbadde86c7f0a1a624f11f887247939e428e378a08ed00913ddc9c7b",
|
| 483 |
+
"src/ouro_tf/hf/config.py": "40f640886aa8dc08995d4c106b6bfa2458f1b2ede8a34607ec73ab7ab6794c66",
|
| 484 |
+
"src/ouro_tf/hf/model.py": "3bdc66c59a83e2d02349d78504f54653de3d76a9655bcda9e7b0c5ed5f662e73",
|
| 485 |
+
"src/ouro_tf/hf/paper_backbone.py": "7ca0d348ace2d4476f840e7a35ae2db6e3ecf8889996dea50fda80060a5ab447",
|
| 486 |
+
"src/ouro_tf/hf/plugin.py": "e63e830f2dbbbcc21ccd0d0373634e8e1a8a4da9d9fb5ec92e25621c8ac88c28",
|
| 487 |
+
"src/ouro_tf/hf/task.py": "4f5ee8a7adc5d227e9ab717e7f930990dfbc86ab56cdb03bad7b55e1fdbe6b62",
|
| 488 |
+
"src/ouro_tf/hf/train.py": "37e775223efd3a87ccf85db312b0e729c793afe42652702acd8a84070bfbe0be",
|
| 489 |
+
"src/ouro_tf/megatron/__init__.py": "33608673413945f504d41a35eec819b32679d95b5440c2f5c4040211b629dd8d",
|
| 490 |
+
"src/ouro_tf/megatron/builder.py": "903ff9c341ac38325465c16edc893cec3eac77546fb2336e5f552cc494fead55",
|
| 491 |
+
"src/ouro_tf/megatron/config.py": "42400d266ff4f931ed714631c7131a38e163dedb58c88079413e066df5b2ec02",
|
| 492 |
+
"src/ouro_tf/megatron/model.py": "a47f9b0323057c2051d7ba3661eab6fcb66ff8299368287d68b38efffb07e01d",
|
| 493 |
+
"src/ouro_tf/megatron/plugin.py": "f503136dcc0f4dc9149b342e63c1294e7f289836630c35a83ccdeff056590127",
|
| 494 |
+
"src/ouro_tf/megatron/task.py": "357d4dfbe54658408237ab4b46b117a70a0cddf9a5139db3cb5e0c9c8220bc2a",
|
| 495 |
+
"src/ouro_tf/megatron/train.py": "094fb39647af251e9064241e9e3cfdf5eead1108e57de00c4f8e9f90da4a501f",
|
| 496 |
+
"src/ouro_tf/policy.py": "431640e74eeadb5061080dc9e8773856d51642c48b77493223e50991137427b5"
|
| 497 |
+
}
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/tokenizer/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
small_search_1epoch/Lookahead-K3-LA2-A2-R0-D0-P0p2-F0-seed0-W760-heads10-H8-L8x3-hrmCycle-HC2-Ref3x2-lowRealFinal-refinerMemoryOnly-AdamATan2-EMA-refinedPseudoFeedback-auxAllPasses-1epoch__20261010-034356_bins16-dp32--r4jyjwuy/step_114274/training_code.tar.gz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e23bf06d79b84d6b355509b3e857928ce842852419f1c1a2a39a6e0ea37669c1
|
| 3 |
+
size 626824
|