diff --git a/.gitattributes b/.gitattributes index 726dc2bf3201d869992c2b2fed0cf7d1628bf9f6..f32f1e39a17d3192d92b7b0b0e99c70322291da0 100644 --- a/.gitattributes +++ b/.gitattributes @@ -60,3 +60,11 @@ mathmedium2-mutualposclip0.5-llama3.2-3b-it-oldreward-4k-corrupt-0.2/global_step mathmedium2-mutualposclip0.0-llama3.2-3b-it-oldreward-4k-corrupt-0.8/global_step_100/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_300/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_250/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_100/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/tokenizer.json filter=lfs diff=lfs merge=lfs -text +pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/tokenizer.json filter=lfs diff=lfs merge=lfs -text +pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/tokenizer.json filter=lfs diff=lfs merge=lfs -text +pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/tokenizer.json filter=lfs diff=lfs merge=lfs -text +pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/tokenizer.json filter=lfs diff=lfs merge=lfs -text +mathhard2-mutualposclip0.0-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json filter=lfs diff=lfs merge=lfs -text +mathhard2-mutualposclip0.2-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json filter=lfs diff=lfs merge=lfs -text diff --git a/mathhard2-mutualposclip0.0-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/model.safetensors b/mathhard2-mutualposclip0.0-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..b3e6af82c5bfa23dfe856271ee4a74cdd6461914 --- /dev/null +++ b/mathhard2-mutualposclip0.0-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:59ea16b12eebddf2473be9c997adcb716220c0cd58aa566f461460189c4f9b5b +size 4063515640 diff --git a/mathhard2-mutualposclip0.0-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json b/mathhard2-mutualposclip0.0-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..cd71f61a15a522601badb3dc960d800d9cb3766c --- /dev/null +++ b/mathhard2-mutualposclip0.0-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4 +size 11422654 diff --git a/mathhard2-mutualposclip0.2-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/model.safetensors b/mathhard2-mutualposclip0.2-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/model.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..073f98db5a23f64a68968e9bc549e51390c08e7e --- /dev/null +++ b/mathhard2-mutualposclip0.2-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/model.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:48130a7669f533aef4cb0e243d22968f3b818a495131b43282920b36a19401dc +size 4063515640 diff --git a/mathhard2-mutualposclip0.2-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json b/mathhard2-mutualposclip0.2-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..cd71f61a15a522601badb3dc960d800d9cb3766c --- /dev/null +++ b/mathhard2-mutualposclip0.2-qwen3-1.7b-oldreward-4k-corrupt-0.2-step-250/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4 +size 11422654 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_0.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_0.pt new file mode 100644 index 0000000000000000000000000000000000000000..db9ff19cf7eb291575cf41fdc349afc01e2ae2ac --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_0.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5c5cbef050241cc58be4305c0b7a6811861e4d4c9f6ea35533d2ae4485d4a180 +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_1.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..a61548dff4f602f418a1e3eeabdd1035f9b32382 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3393a565ce5bf8e8a74875d003a0d1b6c32eace87e742a522af1afbde72dba54 +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_2.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..40835a4c05f7bc78988f1ad19e6e0ce44cce1bf3 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9708ed0f6034ca6315d54f4bf5652ac35f692b344535884605474b8fec3d7364 +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_3.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..79a55c452cfa1b59f93e8272a9859abec1461615 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ddd0ae1e79ef0376767ce85e784a5bd7794ca47e3dc51ab67d31fb554607f740 +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_4.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_4.pt new file mode 100644 index 0000000000000000000000000000000000000000..771a70d240a086e8a5a80a3fa3233946d713aade --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_4.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:184916ee9877fb3b00b03dfe6225617fd5daf19311af081b73819d733246a50d +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_5.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_5.pt new file mode 100644 index 0000000000000000000000000000000000000000..041b4f454bdb35e4bf3abd412a276e7d8187696a --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_5.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:221e913ab4b7193c41c7e820e49c4e47c43c7dae087bc771efcccfc6f21a4f9a +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_6.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_6.pt new file mode 100644 index 0000000000000000000000000000000000000000..7a0456e0f378e6ed568ba5a3067cdcf5e3af7c39 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_6.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:03fc9f4bc4fd8db8c932e599e5795b93e23cef610f73dbc2b0016f7747ddbd2f +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_7.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_7.pt new file mode 100644 index 0000000000000000000000000000000000000000..6d6381738f94910f5851f9b54249f77a260e9ec9 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/extra_state_world_size_8_rank_7.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:476446d582f98cc0c4f84c02ac0feead8d749ff12c5aa8c2ff564200691cbaf0 +size 14632 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_1.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..2b50812e346e56c3379b995433cfecefca2cc08d --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a919122215f4ac0628eeb139dcddd4cd1384cf568e2fa03d37e433762e953ac5 +size 1016051898 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_2.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..876e66b6e8a316c3475682ef3f5acfc78c8a58fc --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9bb9410b4ae6247316ca0e4e62b67ab786e6bfb41df2be18dabf47da9e219c4f +size 1016051898 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_3.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..34aa5718eb75ebfc9a5d3952bcc1e8ded4f1f61a --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6696779968ecc484570516f3e6147be7b547db9ef50622c15c69b4582edd90cc +size 1016051898 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_4.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_4.pt new file mode 100644 index 0000000000000000000000000000000000000000..7594ad435210d3a0e31197c0831c98c45b460d3c --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_4.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:84f8158a93d07842052a75315f1036f00cf0199159030db4cd33af373a45a144 +size 1016051898 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_5.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_5.pt new file mode 100644 index 0000000000000000000000000000000000000000..8e587385d5eb3e264c5e39e91c70c013cfba9144 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_5.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3d99ad08eb7454a321ed98f750bbf6fbec8fd809ef48e53e2e88ce04c2cfbad5 +size 1016051898 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_6.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_6.pt new file mode 100644 index 0000000000000000000000000000000000000000..744fd1071a7a20348bba6cafe587c150fc6062f7 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_6.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:0a77317e54650e111e885f3f532780308c4d43276b67d7b3667a3543d6ebffef +size 1016051898 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_7.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_7.pt new file mode 100644 index 0000000000000000000000000000000000000000..6ce5aa34560bb06fa0e486dfe8e15f4023b30409 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/model_world_size_8_rank_7.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:f39d253268669a11f57da89dcbc14ec4d8c0437e5169df43a4b4416ab4ecf377 +size 1016051898 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_1.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_1.pt new file mode 100644 index 0000000000000000000000000000000000000000..f94c3d795f54113100e81902670f3fb6ae3b5d85 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_1.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3af4b6d5095605167b8122f2f38eb6279314cc9b5e40f0be6e5d333dec398c26 +size 1720600311 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_2.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_2.pt new file mode 100644 index 0000000000000000000000000000000000000000..3fccf1e7ac731bd8fb618e1a911a05370b287ffa --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_2.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7ab5d47e81f85297e667c8d636025d4b2d9743ab3c3b87b57bdc4f3ac23f21ca +size 1720600311 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_3.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_3.pt new file mode 100644 index 0000000000000000000000000000000000000000..801a0f611b48f677f7eee5ef3baaf1f79e95db71 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_3.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:5f214f1c48f7bdf14dbac4c2256c7fee548a8a65c4e0062d60b92c3732784c1c +size 1720600311 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_4.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_4.pt new file mode 100644 index 0000000000000000000000000000000000000000..1bc5566081f5eade50c93b53cc847071d0e8e398 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_4.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:d65b96c00a019fab264b155a77ded801a74de1c9c5bf3de34203a7f78bd8c664 +size 1720600311 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_5.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_5.pt new file mode 100644 index 0000000000000000000000000000000000000000..25d2242b6179801096068a67762fce3f81866639 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_5.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ea37774dfd31861e6af31f6e4d6cb8cca0102c5127ab66a407d2d1212744a41a +size 1720600311 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_6.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_6.pt new file mode 100644 index 0000000000000000000000000000000000000000..f845e2572cda2ba1c5a01f7a546f243884fb6d9c --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_6.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:80b0212e0b4d72d3ef9fed0f81c5fcf2357cdac882aacc57f980787d3b19cd72 +size 1720600311 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_7.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_7.pt new file mode 100644 index 0000000000000000000000000000000000000000..39a49bea42cf1f3d653a28fa7c69f319025acf42 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/actor/optim_world_size_8_rank_7.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:903b636792f0395efd91e1a0ca32c7c065632089a9d2c7179fe6e8c58a30a78f +size 1720600311 diff --git a/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/data.pt b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..ea86e96bfe01f6ee08ce545e9db9a6507ac73db8 --- /dev/null +++ b/mathmedium2-grpo-qwen3-1.7b-4k/global_step_300/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:312d70eb30c6b3a1f6108f7c10d6172a17fd44920565cd983631939c4024504c +size 1492 diff --git a/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_100/actor/huggingface/model-00003-of-00003.safetensors b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_100/actor/huggingface/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c4ef25765ca9b0bf08c8c753261209986a1b218f --- /dev/null +++ b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_100/actor/huggingface/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7d9125b5189fa7a01da8c3721f6cdfc440c3fc9b1bc8f326774887a78184eaf1 +size 4495462048 diff --git a/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_100/actor/huggingface/tokenizer.json b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_100/actor/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2 --- /dev/null +++ b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_100/actor/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00001-of-00003.safetensors b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..abb61bd6eebe8bd8a3f9afb2bfa02b713d174ae8 --- /dev/null +++ b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:608e6e455d018229013244596a38715d4df98a2eef3933824a6c21d1aef996dc +size 4998767360 diff --git a/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00002-of-00003.safetensors b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..c878439cde2de794e83e629c44283dc8edd6afab --- /dev/null +++ b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:71f85d948f8ac074d6949bd0b5e0114e5e058bf2d63cb8ba42e4051b60e898da +size 4932808968 diff --git a/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00003-of-00003.safetensors b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..a6c8591a69cce80fd166c2c0cfd2aed397700f5b --- /dev/null +++ b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:3de9916c5ab2d2c55ae6631e5ba601c5ea0142e30a53717457fd703b4beb1e2c +size 4495462048 diff --git a/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/tokenizer.json b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2 --- /dev/null +++ b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/actor/huggingface/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/data.pt b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/data.pt new file mode 100644 index 0000000000000000000000000000000000000000..bb51ea24dea3b1fe4167726617268b81cb07674d --- /dev/null +++ b/mathmedium2-mutualposclip0.2-llama3.2-3b-it-oldreward-4k-corrupt_delete-0.8/global_step_200/data.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:ccf123f178a85ef6d50972faef86705bf783bcb2ee7b289af10a0b6c2180f9bf +size 1492 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00001-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..97364e41f563c49a6d967b91e30805a59cbe6681 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e4c3ae512f78eda87cf649bd7c783a06d3255ab5c6caded768043e32771ecdbb +size 4998767360 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00002-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..faf49fdea0a9cfe8c4b79ae1013979715e774050 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:61bb19e2687109b0328d7b950817e7493e49a427322611f7c3e84b35ff98ea1c +size 4932808968 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00003-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..80e495fb51d32b3c9fe63aed1b517f1fb771a413 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:86cd49984db16819f43f8ae0302082dfd0822b88c5aa0d0e67738816d15b16fb +size 4495462048 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/tokenizer.json b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/actor/global_step_300/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00001-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..30a2bc3c54e8600f2fdf0dc96dff616f162616ba --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8a45ca86971cc62bcd6be05f49a6fb44ed2bfbf5fefce89f7056af4bec871de +size 4998767360 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00002-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..7676a01f55a4b4ce75e1fdab3ea0cf9409de41ce --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bd401b3e25bc432b2c3d3bdcda7a1a031fbe7dc1b107abfd4ba7374b6697f11b +size 4932808968 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00003-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..e1f92a527e5a5c3e31dbd2eb533f46bc92cddc72 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bc1f0c28855b2da546143765d71be80f8fc98073cbf7ad6b17abacc6a73b6a7b +size 2919464628 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/tokenizer.json b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-em-warmup-0.05-rouge-rouge7-temperature-1.0/critic/global_step_300/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00001-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..42c0dea726adba5fe1b3e3f6b75a2592f973d779 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:bb1e944bbc9ad3730bc0fae46d770eeeff7c739b1322e6738b6b6a86a13c534c +size 4998767360 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00002-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..6ac0d4bf19261b374c247ba21d789bf806c2a4f1 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:29fbb644322cd2a44bd52ae8f6ccee2bb8ff8fc80bb72047fec8cbe4a998cf08 +size 4932808968 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00003-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..3382a3eb2b5f81ed7f37a2e09efca66fe852c513 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:4ec16a4c8bf5338a491c9aacf11f3a1593d6c199a6b707e51eb2e501726569fc +size 4495462048 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/tokenizer.json b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/actor/global_step_300/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00001-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00001-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..2c1fb388173abdd73563ef3f827b70e47ded9550 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00001-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a9ad4422230ca25238231df51c7c6d19a71b7a87dd807f0728f49267237b1448 +size 4998767360 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00002-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00002-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..9e18ef30bab28e5d84931c89a03898ca4e7076f4 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00002-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9458276bbeb2a8d530d62bcd3a209c51be00caeb89b33c08f3fd57f01e105855 +size 4932808968 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00003-of-00003.safetensors b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00003-of-00003.safetensors new file mode 100644 index 0000000000000000000000000000000000000000..88dca52c957db8139400aa0f1210f0229a899030 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/model-00003-of-00003.safetensors @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c9c4f79caf1f8626ce705f0d95fb0f81dc1abb4c55b76d87f92f58f9bc668f69 +size 2919464628 diff --git a/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/tokenizer.json b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/tokenizer.json new file mode 100644 index 0000000000000000000000000000000000000000..1c1d8d5c9024994f1d3b00f9662b8dd89ca13cf2 --- /dev/null +++ b/pilecc_cont1-r1-ppo-llama3.2-3b-it-em-warmup-0.05-rouge-rouge5-temperature-1.0/critic/global_step_300/tokenizer.json @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6b9e4e7fb171f92fd137b777cc2714bf87d11576700a1dcd7a399e7bbe39537b +size 17209920