diff --git a/checkpoint-10/adapter_config.json b/checkpoint-10/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-10/adapter_config.json +++ b/checkpoint-10/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-10/trainer_state.json b/checkpoint-10/trainer_state.json index f3bc7a6c7a8e7a542613193645c83c0135cfaace..d9543360277532e9321b29edfc3279b65f6493cb 100644 --- a/checkpoint-10/trainer_state.json +++ b/checkpoint-10/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 6, "best_metric": 0.012996690347790718, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-6", "epoch": 2.4210526315789473, @@ -7,7 +6,7 @@ "global_step": 10, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 } diff --git a/checkpoint-10/training_args.bin b/checkpoint-10/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-10/training_args.bin +++ b/checkpoint-10/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-12/adapter_config.json b/checkpoint-12/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-12/adapter_config.json +++ b/checkpoint-12/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-12/trainer_state.json b/checkpoint-12/trainer_state.json index 8d028d3771c94b5ea13dd49451668c8949834618..63af88b38df0307c22be3faf58351ed68fccbb37 100644 --- a/checkpoint-12/trainer_state.json +++ b/checkpoint-12/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 12, "best_metric": 0.03234308212995529, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-12", "epoch": 2.8421052631578947, @@ -7,7 +6,7 @@ "global_step": 12, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 diff --git a/checkpoint-12/training_args.bin b/checkpoint-12/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-12/training_args.bin +++ b/checkpoint-12/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-14/adapter_config.json b/checkpoint-14/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-14/adapter_config.json +++ b/checkpoint-14/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-14/trainer_state.json b/checkpoint-14/trainer_state.json index 4f77c8cd47b58975e4fc2300556d90007758118f..0283ff9c783b7aa5ca912544b2dd35fcf23f9a99 100644 --- a/checkpoint-14/trainer_state.json +++ b/checkpoint-14/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 12, "best_metric": 0.03234308212995529, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-12", "epoch": 3.4210526315789473, @@ -7,7 +6,7 @@ "global_step": 14, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 } diff --git a/checkpoint-14/training_args.bin b/checkpoint-14/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-14/training_args.bin +++ b/checkpoint-14/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-16/adapter_config.json b/checkpoint-16/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-16/adapter_config.json +++ b/checkpoint-16/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-16/trainer_state.json b/checkpoint-16/trainer_state.json index 5e83c92bde4761577fc169c57fbd06c509552cad..9227c90abf65fb0517b0d1c7eb78e18f0426365e 100644 --- a/checkpoint-16/trainer_state.json +++ b/checkpoint-16/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 12, "best_metric": 0.03234308212995529, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-12", "epoch": 3.8421052631578947, @@ -7,7 +6,7 @@ "global_step": 16, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 } diff --git a/checkpoint-16/training_args.bin b/checkpoint-16/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-16/training_args.bin +++ b/checkpoint-16/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-18/adapter_config.json b/checkpoint-18/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-18/adapter_config.json +++ b/checkpoint-18/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-18/trainer_state.json b/checkpoint-18/trainer_state.json index a1c1a4655ab577f58a2feda407b25c2a37f8a847..5e552ea3183048034e8097ff86a4f6d1c899830d 100644 --- a/checkpoint-18/trainer_state.json +++ b/checkpoint-18/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 18, "best_metric": 0.03729328140616417, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-18", "epoch": 4.421052631578947, @@ -7,7 +6,7 @@ "global_step": 18, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 }, @@ -269,7 +268,7 @@ "kl": 0.1787109375, "learning_rate": 9.01061596377522e-05, "loss": -0.04504441097378731, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.027318883687257767, "reward_std": 0.10441224090754986, @@ -283,7 +282,7 @@ "grad_norm": 0.005998397711664438, "learning_rate": 8.83022221559489e-05, "loss": -0.045487549155950546, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 18, "train_speed(iter/s)": 0.000432 }, @@ -298,7 +297,7 @@ "eval_reward_std": 0.10691346973180771, "eval_rewards/CosineReward": 0.03729327768087387, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1041.2321, + "eval_runtime": 1041.231, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 18 diff --git a/checkpoint-18/training_args.bin b/checkpoint-18/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-18/training_args.bin +++ b/checkpoint-18/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-2/adapter_config.json b/checkpoint-2/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-2/adapter_config.json +++ b/checkpoint-2/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-2/trainer_state.json b/checkpoint-2/trainer_state.json index f1eb74f328226054adf4f3fb18e6428f343d668d..abb60dd05407f920c1cb0c23f346741d8e8e4205 100644 --- a/checkpoint-2/trainer_state.json +++ b/checkpoint-2/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.42105263157894735, @@ -7,7 +6,7 @@ "global_step": 2, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 } ], "logging_steps": 1, diff --git a/checkpoint-2/training_args.bin b/checkpoint-2/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-2/training_args.bin +++ b/checkpoint-2/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-20/adapter_config.json b/checkpoint-20/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-20/adapter_config.json +++ b/checkpoint-20/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-20/trainer_state.json b/checkpoint-20/trainer_state.json index 799cd565c5d6a9d014372dae066e8e90a4c1adf3..939e242f4ac78f4a828aaf48a8eafdabdd7268dd 100644 --- a/checkpoint-20/trainer_state.json +++ b/checkpoint-20/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 18, "best_metric": 0.03729328140616417, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-18", "epoch": 4.842105263157895, @@ -7,7 +6,7 @@ "global_step": 20, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 }, @@ -269,7 +268,7 @@ "kl": 0.1787109375, "learning_rate": 9.01061596377522e-05, "loss": -0.04504441097378731, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.027318883687257767, "reward_std": 0.10441224090754986, @@ -283,7 +282,7 @@ "grad_norm": 0.005998397711664438, "learning_rate": 8.83022221559489e-05, "loss": -0.045487549155950546, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 18, "train_speed(iter/s)": 0.000432 }, @@ -298,7 +297,7 @@ "eval_reward_std": 0.10691346973180771, "eval_rewards/CosineReward": 0.03729327768087387, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1041.2321, + "eval_runtime": 1041.231, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 18 @@ -311,7 +310,7 @@ "kl": 0.1820068359375, "learning_rate": 8.636868207865244e-05, "loss": -0.03466903418302536, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.466796875, "reward": 0.04069916973821819, "reward_std": 0.11991005763411522, @@ -327,7 +326,7 @@ "kl": 0.19287109375, "learning_rate": 8.43120818934367e-05, "loss": -0.03502114117145538, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 20, "train_speed(iter/s)": 0.000424 } diff --git a/checkpoint-20/training_args.bin b/checkpoint-20/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-20/training_args.bin +++ b/checkpoint-20/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-22/adapter_config.json b/checkpoint-22/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-22/adapter_config.json +++ b/checkpoint-22/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-22/trainer_state.json b/checkpoint-22/trainer_state.json index 38dfc6c5cc5295ebde051bb509900e4bfe319f90..f25a50fd2117664a679744f3e1068c255dd295e1 100644 --- a/checkpoint-22/trainer_state.json +++ b/checkpoint-22/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 18, "best_metric": 0.03729328140616417, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-18", "epoch": 5.421052631578947, @@ -7,7 +6,7 @@ "global_step": 22, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 }, @@ -269,7 +268,7 @@ "kl": 0.1787109375, "learning_rate": 9.01061596377522e-05, "loss": -0.04504441097378731, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.027318883687257767, "reward_std": 0.10441224090754986, @@ -283,7 +282,7 @@ "grad_norm": 0.005998397711664438, "learning_rate": 8.83022221559489e-05, "loss": -0.045487549155950546, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 18, "train_speed(iter/s)": 0.000432 }, @@ -298,7 +297,7 @@ "eval_reward_std": 0.10691346973180771, "eval_rewards/CosineReward": 0.03729327768087387, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1041.2321, + "eval_runtime": 1041.231, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 18 @@ -311,7 +310,7 @@ "kl": 0.1820068359375, "learning_rate": 8.636868207865244e-05, "loss": -0.03466903418302536, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.466796875, "reward": 0.04069916973821819, "reward_std": 0.11991005763411522, @@ -327,7 +326,7 @@ "kl": 0.19287109375, "learning_rate": 8.43120818934367e-05, "loss": -0.03502114117145538, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 20, "train_speed(iter/s)": 0.000424 }, @@ -339,14 +338,14 @@ "kl": 0.17626953125, "learning_rate": 8.213938048432697e-05, "loss": -0.008662773296236992, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.04996980866417289, "reward_std": 0.13849420100450516, "rewards/CosineReward": 0.049969930201768875, "rewards/RepetitionPenalty": -1.1864573679076784e-07, "step": 21, - "train_speed(iter/s)": 0.000407 + "train_speed(iter/s)": 0.000408 }, { "clip_ratio": 5.869188044016482e-05, @@ -355,7 +354,7 @@ "kl": 0.178955078125, "learning_rate": 7.985792958513931e-05, "loss": -0.008743642829358578, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 22, "train_speed(iter/s)": 0.000426 } diff --git a/checkpoint-22/training_args.bin b/checkpoint-22/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-22/training_args.bin +++ b/checkpoint-22/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-24/adapter_config.json b/checkpoint-24/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-24/adapter_config.json +++ b/checkpoint-24/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-24/trainer_state.json b/checkpoint-24/trainer_state.json index 6882365a9f6cb165794d215b4d3e0b03b2c144aa..28c98e01019184c604a3825cec4b497fc6482848 100644 --- a/checkpoint-24/trainer_state.json +++ b/checkpoint-24/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 24, "best_metric": 0.04339282959699631, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-24", "epoch": 5.842105263157895, @@ -7,7 +6,7 @@ "global_step": 24, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 }, @@ -269,7 +268,7 @@ "kl": 0.1787109375, "learning_rate": 9.01061596377522e-05, "loss": -0.04504441097378731, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.027318883687257767, "reward_std": 0.10441224090754986, @@ -283,7 +282,7 @@ "grad_norm": 0.005998397711664438, "learning_rate": 8.83022221559489e-05, "loss": -0.045487549155950546, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 18, "train_speed(iter/s)": 0.000432 }, @@ -298,7 +297,7 @@ "eval_reward_std": 0.10691346973180771, "eval_rewards/CosineReward": 0.03729327768087387, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1041.2321, + "eval_runtime": 1041.231, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 18 @@ -311,7 +310,7 @@ "kl": 0.1820068359375, "learning_rate": 8.636868207865244e-05, "loss": -0.03466903418302536, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.466796875, "reward": 0.04069916973821819, "reward_std": 0.11991005763411522, @@ -327,7 +326,7 @@ "kl": 0.19287109375, "learning_rate": 8.43120818934367e-05, "loss": -0.03502114117145538, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 20, "train_speed(iter/s)": 0.000424 }, @@ -339,14 +338,14 @@ "kl": 0.17626953125, "learning_rate": 8.213938048432697e-05, "loss": -0.008662773296236992, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.04996980866417289, "reward_std": 0.13849420100450516, "rewards/CosineReward": 0.049969930201768875, "rewards/RepetitionPenalty": -1.1864573679076784e-07, "step": 21, - "train_speed(iter/s)": 0.000407 + "train_speed(iter/s)": 0.000408 }, { "clip_ratio": 5.869188044016482e-05, @@ -355,7 +354,7 @@ "kl": 0.178955078125, "learning_rate": 7.985792958513931e-05, "loss": -0.008743642829358578, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 22, "train_speed(iter/s)": 0.000426 }, @@ -367,7 +366,7 @@ "kl": 0.1796875, "learning_rate": 7.74754489035403e-05, "loss": -0.03423420712351799, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.583984375, "reward": 0.034468831261619925, "reward_std": 0.11841745302081108, @@ -381,7 +380,7 @@ "grad_norm": 0.014131724834442139, "learning_rate": 7.500000000000001e-05, "loss": -0.03426633030176163, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 24, "train_speed(iter/s)": 0.000427 }, @@ -396,7 +395,7 @@ "eval_reward_std": 0.10456253588199615, "eval_rewards/CosineReward": 0.04339282959699631, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1045.0642, + "eval_runtime": 1045.0632, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 24 diff --git a/checkpoint-24/training_args.bin b/checkpoint-24/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-24/training_args.bin +++ b/checkpoint-24/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-26/adapter_config.json b/checkpoint-26/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-26/adapter_config.json +++ b/checkpoint-26/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-26/trainer_state.json b/checkpoint-26/trainer_state.json index 879395d69f8fcc796e0f0b4c11b379c5c04ad9cb..5b072b987db923c7897214c7f630cb7ed87c56ab 100644 --- a/checkpoint-26/trainer_state.json +++ b/checkpoint-26/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 24, "best_metric": 0.04339282959699631, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-24", "epoch": 6.421052631578947, @@ -7,7 +6,7 @@ "global_step": 26, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 }, @@ -269,7 +268,7 @@ "kl": 0.1787109375, "learning_rate": 9.01061596377522e-05, "loss": -0.04504441097378731, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.027318883687257767, "reward_std": 0.10441224090754986, @@ -283,7 +282,7 @@ "grad_norm": 0.005998397711664438, "learning_rate": 8.83022221559489e-05, "loss": -0.045487549155950546, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 18, "train_speed(iter/s)": 0.000432 }, @@ -298,7 +297,7 @@ "eval_reward_std": 0.10691346973180771, "eval_rewards/CosineReward": 0.03729327768087387, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1041.2321, + "eval_runtime": 1041.231, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 18 @@ -311,7 +310,7 @@ "kl": 0.1820068359375, "learning_rate": 8.636868207865244e-05, "loss": -0.03466903418302536, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.466796875, "reward": 0.04069916973821819, "reward_std": 0.11991005763411522, @@ -327,7 +326,7 @@ "kl": 0.19287109375, "learning_rate": 8.43120818934367e-05, "loss": -0.03502114117145538, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 20, "train_speed(iter/s)": 0.000424 }, @@ -339,14 +338,14 @@ "kl": 0.17626953125, "learning_rate": 8.213938048432697e-05, "loss": -0.008662773296236992, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.04996980866417289, "reward_std": 0.13849420100450516, "rewards/CosineReward": 0.049969930201768875, "rewards/RepetitionPenalty": -1.1864573679076784e-07, "step": 21, - "train_speed(iter/s)": 0.000407 + "train_speed(iter/s)": 0.000408 }, { "clip_ratio": 5.869188044016482e-05, @@ -355,7 +354,7 @@ "kl": 0.178955078125, "learning_rate": 7.985792958513931e-05, "loss": -0.008743642829358578, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 22, "train_speed(iter/s)": 0.000426 }, @@ -367,7 +366,7 @@ "kl": 0.1796875, "learning_rate": 7.74754489035403e-05, "loss": -0.03423420712351799, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.583984375, "reward": 0.034468831261619925, "reward_std": 0.11841745302081108, @@ -381,7 +380,7 @@ "grad_norm": 0.014131724834442139, "learning_rate": 7.500000000000001e-05, "loss": -0.03426633030176163, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 24, "train_speed(iter/s)": 0.000427 }, @@ -396,7 +395,7 @@ "eval_reward_std": 0.10456253588199615, "eval_rewards/CosineReward": 0.04339282959699631, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1045.0642, + "eval_runtime": 1045.0632, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 24 @@ -409,7 +408,7 @@ "kl": 0.1800537109375, "learning_rate": 7.243995901002312e-05, "loss": -0.02097315341234207, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.6171875, "reward": 0.03010205877944827, "reward_std": 0.10742511600255966, @@ -425,7 +424,7 @@ "kl": 0.18408203125, "learning_rate": 6.980398830195785e-05, "loss": -0.02103913575410843, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 26, "train_speed(iter/s)": 0.000421 } diff --git a/checkpoint-26/training_args.bin b/checkpoint-26/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-26/training_args.bin +++ b/checkpoint-26/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-28/adapter_config.json b/checkpoint-28/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-28/adapter_config.json +++ b/checkpoint-28/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-28/trainer_state.json b/checkpoint-28/trainer_state.json index 54f298b29417457381943a331bc7ae1242c31d16..148e40ed0b36a607a26cd48bcc7a44e47f551f9b 100644 --- a/checkpoint-28/trainer_state.json +++ b/checkpoint-28/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 24, "best_metric": 0.04339282959699631, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-24", "epoch": 6.842105263157895, @@ -7,7 +6,7 @@ "global_step": 28, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 }, @@ -269,7 +268,7 @@ "kl": 0.1787109375, "learning_rate": 9.01061596377522e-05, "loss": -0.04504441097378731, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.027318883687257767, "reward_std": 0.10441224090754986, @@ -283,7 +282,7 @@ "grad_norm": 0.005998397711664438, "learning_rate": 8.83022221559489e-05, "loss": -0.045487549155950546, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 18, "train_speed(iter/s)": 0.000432 }, @@ -298,7 +297,7 @@ "eval_reward_std": 0.10691346973180771, "eval_rewards/CosineReward": 0.03729327768087387, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1041.2321, + "eval_runtime": 1041.231, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 18 @@ -311,7 +310,7 @@ "kl": 0.1820068359375, "learning_rate": 8.636868207865244e-05, "loss": -0.03466903418302536, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.466796875, "reward": 0.04069916973821819, "reward_std": 0.11991005763411522, @@ -327,7 +326,7 @@ "kl": 0.19287109375, "learning_rate": 8.43120818934367e-05, "loss": -0.03502114117145538, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 20, "train_speed(iter/s)": 0.000424 }, @@ -339,14 +338,14 @@ "kl": 0.17626953125, "learning_rate": 8.213938048432697e-05, "loss": -0.008662773296236992, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.04996980866417289, "reward_std": 0.13849420100450516, "rewards/CosineReward": 0.049969930201768875, "rewards/RepetitionPenalty": -1.1864573679076784e-07, "step": 21, - "train_speed(iter/s)": 0.000407 + "train_speed(iter/s)": 0.000408 }, { "clip_ratio": 5.869188044016482e-05, @@ -355,7 +354,7 @@ "kl": 0.178955078125, "learning_rate": 7.985792958513931e-05, "loss": -0.008743642829358578, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 22, "train_speed(iter/s)": 0.000426 }, @@ -367,7 +366,7 @@ "kl": 0.1796875, "learning_rate": 7.74754489035403e-05, "loss": -0.03423420712351799, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.583984375, "reward": 0.034468831261619925, "reward_std": 0.11841745302081108, @@ -381,7 +380,7 @@ "grad_norm": 0.014131724834442139, "learning_rate": 7.500000000000001e-05, "loss": -0.03426633030176163, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 24, "train_speed(iter/s)": 0.000427 }, @@ -396,7 +395,7 @@ "eval_reward_std": 0.10456253588199615, "eval_rewards/CosineReward": 0.04339282959699631, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1045.0642, + "eval_runtime": 1045.0632, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 24 @@ -409,7 +408,7 @@ "kl": 0.1800537109375, "learning_rate": 7.243995901002312e-05, "loss": -0.02097315341234207, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.6171875, "reward": 0.03010205877944827, "reward_std": 0.10742511600255966, @@ -425,7 +424,7 @@ "kl": 0.18408203125, "learning_rate": 6.980398830195785e-05, "loss": -0.02103913575410843, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 26, "train_speed(iter/s)": 0.000421 }, @@ -437,7 +436,7 @@ "kl": 0.174560546875, "learning_rate": 6.710100716628344e-05, "loss": -0.03593946248292923, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.513671875, "reward": 0.04752760287374258, "reward_std": 0.14935147762298584, @@ -453,7 +452,7 @@ "kl": 0.182373046875, "learning_rate": 6.434016163555452e-05, "loss": -0.03595500811934471, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 28, "train_speed(iter/s)": 0.000422 } diff --git a/checkpoint-28/training_args.bin b/checkpoint-28/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-28/training_args.bin +++ b/checkpoint-28/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-30/adapter_config.json b/checkpoint-30/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-30/adapter_config.json +++ b/checkpoint-30/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-30/args.json b/checkpoint-30/args.json new file mode 100644 index 0000000000000000000000000000000000000000..e5fddc18fe4219cc621c4285a0faf4591e553d74 --- /dev/null +++ b/checkpoint-30/args.json @@ -0,0 +1,416 @@ +{ + "model": "/mnt/nvme5n1p1/distill-14b-rl-70", + "model_type": "deepseek_r1_distill", + "model_revision": null, + "task_type": "causal_lm", + "torch_dtype": "bfloat16", + "attn_impl": null, + "num_labels": null, + "rope_scaling": null, + "device_map": null, + "max_memory": {}, + "local_repo_path": null, + "template": "deepseek_r1", + "system": "You are a helpful and harmless assistant. You are Qwen developed by Alibaba. You should think step-by-step. Return final answer within \\\\boxed{}.", + "max_length": 16384, + "truncation_strategy": "left", + "max_pixels": null, + "tools_prompt": "react_en", + "norm_bbox": null, + "response_prefix": null, + "padding_side": "right", + "loss_scale": "last_round", + "sequence_parallel_size": 1, + "use_chat_template": true, + "template_backend": "swift", + "dataset": [ + "stage2_aime.jsonl" + ], + "val_dataset": [], + "split_dataset_ratio": 0.01, + "data_seed": 42, + "dataset_num_proc": 52, + "streaming": false, + "enable_cache": false, + "download_mode": "reuse_dataset_if_exists", + "columns": {}, + "strict": false, + "remove_unused_columns": false, + "model_name": [ + null, + null + ], + "model_author": [ + null, + null + ], + "custom_dataset_info": [], + "quant_method": null, + "quant_bits": null, + "hqq_axis": null, + "bnb_4bit_compute_dtype": "bfloat16", + "bnb_4bit_quant_type": "nf4", + "bnb_4bit_use_double_quant": true, + "bnb_4bit_quant_storage": null, + "max_new_tokens": 64, + "temperature": 1.0, + "top_k": 50, + "top_p": 0.9, + "repetition_penalty": 1.1, + "num_beams": 1, + "stream": false, + "stop_words": [], + "logprobs": false, + "top_logprobs": null, + "ckpt_dir": "/mnt/nvme5n1p1/distill-14b-rl-70", + "load_dataset_config": null, + "lora_modules": [], + "tuner_backend": "peft", + "train_type": "lora", + "adapters": [], + "external_plugins": [], + "seed": 42, + "model_kwargs": {}, + "load_args": false, + "load_data_args": false, + "use_hf": true, + "hub_token": null, + "custom_register_path": [], + "ignore_args_error": false, + "use_swift_lora": false, + "output_dir": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345", + "overwrite_output_dir": false, + "do_train": false, + "do_eval": false, + "do_predict": false, + "eval_strategy": "steps", + "prediction_loss_only": false, + "per_device_train_batch_size": 4, + "per_device_eval_batch_size": 4, + "per_gpu_train_batch_size": null, + "per_gpu_eval_batch_size": null, + "gradient_accumulation_steps": 4, + "eval_accumulation_steps": null, + "eval_delay": 0, + "torch_empty_cache_steps": null, + "learning_rate": 0.0001, + "weight_decay": 0.1, + "adam_beta1": 0.9, + "adam_beta2": 0.999, + "adam_epsilon": 1e-08, + "max_grad_norm": 1.0, + "num_train_epochs": 15.0, + "max_steps": -1, + "lr_scheduler_type": "cosine", + "lr_scheduler_kwargs": null, + "warmup_ratio": 0.1, + "warmup_steps": 0, + "log_level": "passive", + "log_level_replica": "warning", + "log_on_each_node": true, + "logging_dir": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/runs", + "logging_strategy": "steps", + "logging_first_step": true, + "logging_steps": 1, + "logging_nan_inf_filter": true, + "save_strategy": "steps", + "save_steps": 2.0, + "save_total_limit": 100, + "save_safetensors": true, + "save_on_each_node": true, + "save_only_model": false, + "restore_callback_states_from_checkpoint": false, + "no_cuda": false, + "use_cpu": false, + "use_mps_device": false, + "jit_mode_eval": false, + "use_ipex": false, + "bf16": true, + "fp16": false, + "fp16_opt_level": "O1", + "half_precision_backend": "auto", + "bf16_full_eval": false, + "fp16_full_eval": false, + "tf32": null, + "local_rank": 0, + "ddp_backend": null, + "tpu_num_cores": null, + "tpu_metrics_debug": false, + "debug": null, + "dataloader_drop_last": false, + "eval_steps": 6.0, + "dataloader_num_workers": 52, + "dataloader_prefetch_factor": null, + "past_index": -1, + "run_name": null, + "disable_tqdm": null, + "label_names": null, + "load_best_model_at_end": false, + "metric_for_best_model": "reward", + "greater_is_better": true, + "ignore_data_skip": false, + "fsdp": "", + "fsdp_min_num_params": 0, + "fsdp_config": null, + "fsdp_transformer_layer_cls_to_wrap": null, + "accelerator_config": { + "dispatch_batches": false + }, + "deepspeed": { + "fp16": { + "enabled": "auto", + "loss_scale": 0, + "loss_scale_window": 1000, + "initial_scale_power": 16, + "hysteresis": 2, + "min_loss_scale": 1 + }, + "bf16": { + "enabled": "auto" + }, + "zero_optimization": { + "stage": 2, + "offload_optimizer": { + "device": "none", + "pin_memory": true + }, + "allgather_partitions": true, + "allgather_bucket_size": 200000000.0, + "overlap_comm": true, + "reduce_scatter": true, + "reduce_bucket_size": 200000000.0, + "contiguous_gradients": true + }, + "gradient_accumulation_steps": "auto", + "gradient_clipping": "auto", + "steps_per_print": 2000, + "train_batch_size": "auto", + "train_micro_batch_size_per_gpu": "auto", + "wall_clock_breakdown": false + }, + "label_smoothing_factor": 0.0, + "optim": "adamw_torch", + "optim_args": null, + "adafactor": false, + "group_by_length": false, + "length_column_name": "length", + "report_to": [ + "wandb" + ], + "ddp_find_unused_parameters": null, + "ddp_bucket_cap_mb": null, + "ddp_broadcast_buffers": null, + "dataloader_pin_memory": true, + "dataloader_persistent_workers": false, + "skip_memory_metrics": true, + "use_legacy_prediction_loop": false, + "push_to_hub": false, + "resume_from_checkpoint": null, + "hub_model_id": null, + "hub_strategy": "every_save", + "hub_private_repo": null, + "hub_always_push": false, + "gradient_checkpointing": true, + "gradient_checkpointing_kwargs": null, + "include_inputs_for_metrics": false, + "include_for_metrics": [], + "eval_do_concat_batches": true, + "fp16_backend": "auto", + "evaluation_strategy": "steps", + "push_to_hub_model_id": null, + "push_to_hub_organization": null, + "push_to_hub_token": null, + "mp_parameters": "", + "auto_find_batch_size": false, + "full_determinism": false, + "torchdynamo": null, + "ray_scope": "last", + "ddp_timeout": 1800, + "torch_compile": false, + "torch_compile_backend": null, + "torch_compile_mode": null, + "dispatch_batches": null, + "split_batches": null, + "include_tokens_per_second": false, + "include_num_input_tokens_seen": false, + "neftune_noise_alpha": null, + "optim_target_modules": null, + "batch_eval_metrics": false, + "eval_on_start": false, + "use_liger_kernel": false, + "eval_use_gather_object": false, + "average_tokens_across_devices": false, + "sortish_sampler": false, + "predict_with_generate": false, + "generation_max_length": null, + "generation_num_beams": null, + "generation_config": null, + "freeze_parameters": [], + "freeze_parameters_ratio": 0.0, + "trainable_parameters": [], + "freeze_llm": false, + "freeze_vit": true, + "freeze_aligner": true, + "target_modules": [ + "all-linear" + ], + "target_regex": null, + "modules_to_save": [], + "lora_rank": 32, + "lora_alpha": 32, + "lora_dropout": 0.05, + "lora_bias": "none", + "lora_dtype": null, + "lorap_lr_ratio": null, + "use_rslora": false, + "use_dora": false, + "lora_ga_batch_size": 2, + "lora_ga_iters": 2, + "lora_ga_max_length": 1024, + "lora_ga_direction": "ArB2r", + "lora_ga_scale": "stable", + "lora_ga_stable_gamma": 16, + "init_weights": true, + "fourier_n_frequency": 2000, + "fourier_scaling": 300.0, + "boft_block_size": 4, + "boft_block_num": 0, + "boft_n_butterfly_factor": 1, + "boft_dropout": 0.0, + "vera_rank": 256, + "vera_projection_prng_key": 0, + "vera_dropout": 0.0, + "vera_d_initial": 0.1, + "adapter_act": "gelu", + "adapter_length": 128, + "use_galore": false, + "galore_target_modules": null, + "galore_rank": 128, + "galore_update_proj_gap": 50, + "galore_scale": 1.0, + "galore_proj_type": "std", + "galore_optim_per_parameter": false, + "galore_with_embedding": false, + "galore_quantization": false, + "galore_proj_quant": false, + "galore_proj_bits": 4, + "galore_proj_group_size": 256, + "galore_cos_threshold": 0.4, + "galore_gamma_proj": 2, + "galore_queue_size": 5, + "adalora_target_r": 8, + "adalora_init_r": 12, + "adalora_tinit": 0, + "adalora_tfinal": 0, + "adalora_deltaT": 1, + "adalora_beta1": 0.85, + "adalora_beta2": 0.85, + "adalora_orth_reg_weight": 0.5, + "llamapro_num_new_blocks": 4, + "llamapro_num_groups": null, + "lisa_activated_layers": 0, + "lisa_step_interval": 20, + "reft_layer_key": null, + "reft_layers": null, + "reft_rank": 4, + "reft_intervention_type": "LoreftIntervention", + "reft_args": null, + "use_liger": false, + "model_layer_cls_name": null, + "metric_warmup_step": 0, + "fsdp_num": 1, + "acc_steps": 1, + "swanlab_token": null, + "swanlab_project": null, + "swanlab_workspace": null, + "swanlab_exp_name": null, + "swanlab_mode": "cloud", + "add_version": true, + "resume_only_model": false, + "check_model": true, + "create_checkpoint_symlink": false, + "packing": false, + "lazy_tokenize": false, + "loss_type": null, + "optimizer": null, + "metric": null, + "acc_strategy": "token", + "zero_hpz_partition_size": null, + "reward_model": null, + "reward_adapters": [], + "reward_model_type": null, + "reward_model_revision": null, + "num_ppo_epochs": 4, + "whiten_rewards": false, + "kl_coef": 0.05, + "cliprange": 0.2, + "vf_coef": 0.1, + "cliprange_value": 0.2, + "gamma": 1.0, + "lam": 0.95, + "num_mini_batches": 1, + "local_rollout_forward_batch_size": 64, + "num_sample_generations": 10, + "response_length": 512, + "missing_eos_penalty": null, + "num_infer_workers": 8, + "vllm_max_num_seqs": 256, + "vllm_enforce_eager": false, + "vllm_limit_mm_per_prompt": null, + "vllm_enable_prefix_caching": true, + "cosine_min_len_value_wrong": 0.0, + "cosine_max_len_value_wrong": -0.2, + "cosine_min_len_value_correct": 0.8, + "cosine_max_len_value_correct": 0.4, + "cosine_max_len": 12288, + "repetition_n_grams": 40, + "repetition_max_penalty": -0.05, + "use_lmdeploy": false, + "lmdeploy_device": "auto", + "lmdeploy_session_len": null, + "lmdeploy_cache_max_entry_count": 0.8, + "async_generate": false, + "tensor_parallel_size": 1, + "sleep_level": 0, + "move_model_batches": null, + "offload_optimizer": false, + "offload_model": false, + "gc_collect_after_offload": false, + "num_generations": 8, + "max_completion_length": 12288, + "ds3_gather_for_generation": true, + "reward_funcs": [ + "cosine", + "repetition" + ], + "reward_weights": null, + "log_completions": true, + "use_vllm": false, + "vllm_device": [ + "auto" + ], + "vllm_gpu_memory_utilization": 0.9, + "vllm_max_model_len": null, + "num_iterations": 2, + "epsilon": 0.2, + "rlhf_type": "grpo", + "ref_model": null, + "ref_model_type": null, + "ref_model_revision": null, + "beta": 0.04, + "label_smoothing": 0, + "rpo_alpha": 1.0, + "cpo_alpha": 1.0, + "simpo_gamma": 1, + "desirable_weight": 1.0, + "undesirable_weight": 1.0, + "rank": 0, + "global_world_size": 32, + "local_world_size": 8, + "model_suffix": "distill-14b-rl-70", + "model_info": "ModelInfo(model_type='deepseek_r1_distill', model_dir='/mnt/nvme5n1p1/distill-14b-rl-70', torch_dtype=torch.bfloat16, max_model_len=131072, quant_method=None, quant_bits=None, rope_scaling=None, config=None, task_type='causal_lm', num_labels=None)", + "model_meta": "ModelMeta(model_type='deepseek_r1_distill', model_groups=[ModelGroup(models=[Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-7B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-7B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-14B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-14B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-32B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Qwen-32B', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=['transformers>=4.37'], tags=[]), ModelGroup(models=[Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-8B', model_path=None, ms_revision=None, hf_revision=None), Model(ms_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-70B', hf_model_id='deepseek-ai/DeepSeek-R1-Distill-Llama-70B', model_path=None, ms_revision=None, hf_revision=None)], ignore_patterns=None, requires=None, tags=[])], template='deepseek_r1', get_function=, model_arch='llama', architectures=['Qwen2ForCausalLM', 'LlamaForCausalLM'], additional_saved_files=[], torch_dtype=None, is_multimodal=False, is_reward=False, task_type=None, ignore_patterns=[], requires=[], tags=[])", + "model_dir": "/mnt/nvme5n1p1/distill-14b-rl-70", + "hub": "", + "training_args": "GRPOConfig(output_dir='/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345', overwrite_output_dir=False, do_train=False, do_eval=True, do_predict=False, eval_strategy=, prediction_loss_only=False, per_device_train_batch_size=4, per_device_eval_batch_size=4, per_gpu_train_batch_size=None, per_gpu_eval_batch_size=None, gradient_accumulation_steps=4, eval_accumulation_steps=None, eval_delay=0, torch_empty_cache_steps=None, learning_rate=0.0001, weight_decay=0.1, adam_beta1=0.9, adam_beta2=0.999, adam_epsilon=1e-08, max_grad_norm=1.0, num_train_epochs=15.0, max_steps=-1, lr_scheduler_type=, lr_scheduler_kwargs=None, warmup_ratio=0.1, warmup_steps=0, log_level='passive', log_level_replica='warning', log_on_each_node=True, logging_dir='/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/runs', logging_strategy=, logging_first_step=True, logging_steps=1, logging_nan_inf_filter=True, save_strategy=, save_steps=2, save_total_limit=100, save_safetensors=True, save_on_each_node=True, save_only_model=False, restore_callback_states_from_checkpoint=False, no_cuda=False, use_cpu=False, use_mps_device=False, seed=42, data_seed=42, jit_mode_eval=False, use_ipex=False, bf16=True, fp16=False, fp16_opt_level='O1', half_precision_backend='auto', bf16_full_eval=False, fp16_full_eval=False, tf32=None, local_rank=0, ddp_backend=None, tpu_num_cores=None, tpu_metrics_debug=False, debug=[], dataloader_drop_last=False, eval_steps=6, dataloader_num_workers=52, dataloader_prefetch_factor=None, past_index=-1, run_name='/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345', disable_tqdm=False, remove_unused_columns=False, label_names=None, load_best_model_at_end=False, metric_for_best_model='reward', greater_is_better=True, ignore_data_skip=False, fsdp=[], fsdp_min_num_params=0, fsdp_config={'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}, fsdp_transformer_layer_cls_to_wrap=None, accelerator_config=AcceleratorConfig(split_batches=False, dispatch_batches=False, even_batches=True, use_seedable_sampler=True, non_blocking=False, gradient_accumulation_kwargs=None, use_configured_state=False), deepspeed={'fp16': {'enabled': 'auto', 'loss_scale': 0, 'loss_scale_window': 1000, 'initial_scale_power': 16, 'hysteresis': 2, 'min_loss_scale': 1}, 'bf16': {'enabled': 'auto'}, 'zero_optimization': {'stage': 2, 'offload_optimizer': {'device': 'none', 'pin_memory': True}, 'allgather_partitions': True, 'allgather_bucket_size': 200000000.0, 'overlap_comm': True, 'reduce_scatter': True, 'reduce_bucket_size': 200000000.0, 'contiguous_gradients': True}, 'gradient_accumulation_steps': 'auto', 'gradient_clipping': 'auto', 'steps_per_print': 2000, 'train_batch_size': 'auto', 'train_micro_batch_size_per_gpu': 'auto', 'wall_clock_breakdown': False}, label_smoothing_factor=0.0, optim=, optim_args=None, adafactor=False, group_by_length=False, length_column_name='length', report_to=['wandb'], ddp_find_unused_parameters=None, ddp_bucket_cap_mb=None, ddp_broadcast_buffers=None, dataloader_pin_memory=True, dataloader_persistent_workers=False, skip_memory_metrics=True, use_legacy_prediction_loop=False, push_to_hub=False, resume_from_checkpoint=None, hub_model_id=None, hub_strategy=, hub_token=None, hub_private_repo=None, hub_always_push=False, gradient_checkpointing=True, gradient_checkpointing_kwargs=None, include_inputs_for_metrics=False, include_for_metrics=[], eval_do_concat_batches=True, fp16_backend='auto', evaluation_strategy='steps', push_to_hub_model_id=None, push_to_hub_organization=None, push_to_hub_token=None, mp_parameters='', auto_find_batch_size=False, full_determinism=False, torchdynamo=None, ray_scope='last', ddp_timeout=1800, torch_compile=False, torch_compile_backend=None, torch_compile_mode=None, dispatch_batches=None, split_batches=None, include_tokens_per_second=None, include_num_input_tokens_seen=None, neftune_noise_alpha=None, optim_target_modules=None, batch_eval_metrics=False, eval_on_start=False, use_liger_kernel=False, eval_use_gather_object=False, average_tokens_across_devices=None, model_init_kwargs=None, max_prompt_length=512, num_generations=8, max_completion_length=12288, ds3_gather_for_generation=True, temperature=1.0, top_p=0.9, top_k=50, min_p=None, repetition_penalty=1.1, cache_implementation=None, use_vllm=False, vllm_server_host='0.0.0.0', vllm_server_port=8000, vllm_server_timeout=120.0, vllm_guided_decoding_regex=None, beta=0.04, num_iterations=2, epsilon=0.2, epsilon_high=None, reward_weights=None, scale_rewards=True, sync_ref_model=False, ref_model_mixup_alpha=0.6, ref_model_sync_steps=512, log_completions=True, vllm_device=['auto'], vllm_gpu_memory_utilization=0.9, vllm_dtype=None, vllm_max_model_len=None, vllm_enable_prefix_caching=True, acc_strategy='token', sequence_parallel_size=1, check_model=True, train_sampler_random=True, is_encoder_decoder=False, metric_warmup_step=0, train_dataset_sample=-1, fsdp_num=1, acc_steps=1, train_type='lora', optimizer=None, local_repo_path=None, galore_config=None, num_infer_workers=8, vllm_max_num_seqs=256, vllm_enforce_eager=False, vllm_limit_mm_per_prompt={}, cosine_min_len_value_wrong=0.0, cosine_max_len_value_wrong=-0.2, cosine_min_len_value_correct=0.8, cosine_max_len_value_correct=0.4, cosine_max_len=12288, repetition_n_grams=40, repetition_max_penalty=-0.05, use_lmdeploy=False, lmdeploy_device='auto', lmdeploy_session_len=None, lmdeploy_cache_max_entry_count=0.8, async_generate=False, tensor_parallel_size=1, sleep_level=0, move_model_batches=None, offload_optimizer=False, offload_model=False, gc_collect_after_offload=False, stop_words=[])" +} \ No newline at end of file diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..8d360ab41641272dfc409b86d3e927eef9c5d8bc --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:116fd5b5622dc5ab41c2b4b69599992b88a191c17075a38cd7dadf69575e4a78 +size 51616517 diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..430cee3f4886efe9cde3275c11b14e524f16173c --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:10e8a2acf8c0ac1a9ffa239a43c6bb598040c0220e25be565e2bb53eaa1bd714 +size 51616005 diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..b452ccbfc8fa394c0437b40e6a67255f1f3ca0ae --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:30af77056d615b52d15c56ee9753e492629de7963f1d6ee5ef8a338886879edb +size 51616517 diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..317c478eb7eb35e05fd1d49fe9ef87582b6261bc --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:9ec075b7a811dd204fd12c6ac145a9dceccdc4f56ffc59571a349c2da1de140f +size 51616005 diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..92f2db55f5cef7972d3dc0522e3b7bcb7852e142 --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:eaec79a6883ccbbfee57c719f45ee94fc319c6a85cf44e731a557599e99e08af +size 51616517 diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..f7685f483c51cfa24d1fa07cc790979280fc8f0a --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:70645416a07a17e9a5d87ba5a084531ba3a59ee64549321dce7acc6591bd3cbd +size 51616005 diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..59c3cfe02b5b5d6685044331db92614c08d49cee --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:dfe1c45f684c024d69a085a30832b079cedf161331d4d1d2b0ca34e083764488 +size 51616517 diff --git a/checkpoint-30/global_step30/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt b/checkpoint-30/global_step30/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt new file mode 100644 index 0000000000000000000000000000000000000000..ab459f40f48e25892ac5f2a64de63841f087205e --- /dev/null +++ b/checkpoint-30/global_step30/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:81e017409e9f8dd518dadb52434224c2a69a811e9ab613d40be8809afe1b1272 +size 51616005 diff --git a/checkpoint-30/rng_state_0.pth b/checkpoint-30/rng_state_0.pth new file mode 100644 index 0000000000000000000000000000000000000000..5c871b2bae16013f4782b467a12b534f33b66baa --- /dev/null +++ b/checkpoint-30/rng_state_0.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:7445d8a3801c08606f052af6eba1f066506146bc8850e86b93b99b9724d6a06b +size 16389 diff --git a/checkpoint-30/rng_state_1.pth b/checkpoint-30/rng_state_1.pth new file mode 100644 index 0000000000000000000000000000000000000000..4942aba1147778c51f49b7f0dc7acc08584c36fc --- /dev/null +++ b/checkpoint-30/rng_state_1.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:babfc1ea5e5f1bc14b7cccbd566cff5d7cb473de1e0d77762bca88dc3bc2dea3 +size 16389 diff --git a/checkpoint-30/rng_state_2.pth b/checkpoint-30/rng_state_2.pth new file mode 100644 index 0000000000000000000000000000000000000000..25bbe1dc3ad6e703ec9341714bd7e0c60904cfad --- /dev/null +++ b/checkpoint-30/rng_state_2.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:a431c5a1142e39a8f6ceca0202b05dafd58fc61cfbf3f69fe6950fa8d0b864ea +size 16389 diff --git a/checkpoint-30/rng_state_3.pth b/checkpoint-30/rng_state_3.pth new file mode 100644 index 0000000000000000000000000000000000000000..1529576376e1f638c1dc09fea31a4775d1a4f3b5 --- /dev/null +++ b/checkpoint-30/rng_state_3.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:c134753e443e6f990e92a23324f4b3e5487167a3cc2777f478d832bd3af0b587 +size 16389 diff --git a/checkpoint-30/rng_state_4.pth b/checkpoint-30/rng_state_4.pth new file mode 100644 index 0000000000000000000000000000000000000000..97f0a435d24893df19013e163e0f3f1009ef839f --- /dev/null +++ b/checkpoint-30/rng_state_4.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e26df1338c58354b929a3988f94ed860e003fcf9b7a805288ea975992aeb1387 +size 16389 diff --git a/checkpoint-30/rng_state_5.pth b/checkpoint-30/rng_state_5.pth new file mode 100644 index 0000000000000000000000000000000000000000..958fb79c1ec1add7a8cef9c1fc39b48e981b5313 --- /dev/null +++ b/checkpoint-30/rng_state_5.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:6208c597b8eaeb85184446a7e8fe032bef0bbfdd8a59aa7c20b1bc8e8b007d34 +size 16325 diff --git a/checkpoint-30/rng_state_6.pth b/checkpoint-30/rng_state_6.pth new file mode 100644 index 0000000000000000000000000000000000000000..e99d9ca2d7ea8a0b7c2afbd2a3413de206ce753b --- /dev/null +++ b/checkpoint-30/rng_state_6.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:e8803b97be66ed6f3d4987aa0c579ab126c1b7e2347e0dcb094f20dcdccebce5 +size 16389 diff --git a/checkpoint-30/rng_state_7.pth b/checkpoint-30/rng_state_7.pth new file mode 100644 index 0000000000000000000000000000000000000000..e4af085f612f2f5a8363fcfb0c20b778fb2eda30 --- /dev/null +++ b/checkpoint-30/rng_state_7.pth @@ -0,0 +1,3 @@ +version https://git-lfs.github.com/spec/v1 +oid sha256:308d4d3ce7e2d5022ad4979d52690b872fca1a21658f154aa030b374f60b565d +size 16453 diff --git a/checkpoint-30/trainer_state.json b/checkpoint-30/trainer_state.json index 664690c82b357f0d4097c0a552d4c28eca3f4ebc..535f7d018c6b69c58687ec47fad131e8d13e5836 100644 --- a/checkpoint-30/trainer_state.json +++ b/checkpoint-30/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 30, "best_metric": 0.05227778106927872, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-30", "epoch": 7.421052631578947, @@ -7,7 +6,7 @@ "global_step": 30, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 }, @@ -143,7 +142,7 @@ "kl": 0.0963134765625, "learning_rate": 9.924038765061042e-05, "loss": -0.05842069163918495, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.255859375, "reward": 0.03643610421568155, "reward_std": 0.11898956261575222, @@ -159,7 +158,7 @@ "kl": 0.1185302734375, "learning_rate": 9.865224352899119e-05, "loss": -0.06491819024085999, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 10, "train_speed(iter/s)": 0.000436 }, @@ -171,7 +170,7 @@ "kl": 0.1275634765625, "learning_rate": 9.789947561577445e-05, "loss": -0.04600231721997261, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.361328125, "reward": 0.023204635945148766, "reward_std": 0.10593634657561779, @@ -185,7 +184,7 @@ "grad_norm": 0.05781339108943939, "learning_rate": 9.698463103929542e-05, "loss": -0.05069056898355484, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 12, "train_speed(iter/s)": 0.000439 }, @@ -200,7 +199,7 @@ "eval_reward_std": 0.10685288906097412, "eval_rewards/CosineReward": 0.03234308212995529, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1025.9048, + "eval_runtime": 1025.9041, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 12 @@ -213,7 +212,7 @@ "kl": 0.151123046875, "learning_rate": 9.591080534401371e-05, "loss": -0.02191038429737091, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.419921875, "reward": 0.035983758978545666, "reward_std": 0.11553369648754597, @@ -229,7 +228,7 @@ "kl": 0.169189453125, "learning_rate": 9.468163201617062e-05, "loss": -0.022672578692436218, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 14, "train_speed(iter/s)": 0.000427 }, @@ -241,7 +240,7 @@ "kl": 0.166748046875, "learning_rate": 9.330127018922194e-05, "loss": -0.059799157083034515, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.4765625, "reward": 0.03584331553429365, "reward_std": 0.11829411797225475, @@ -257,7 +256,7 @@ "kl": 0.16748046875, "learning_rate": 9.177439057064683e-05, "loss": -0.06071458384394646, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 16, "train_speed(iter/s)": 0.000431 }, @@ -269,7 +268,7 @@ "kl": 0.1787109375, "learning_rate": 9.01061596377522e-05, "loss": -0.04504441097378731, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.027318883687257767, "reward_std": 0.10441224090754986, @@ -283,7 +282,7 @@ "grad_norm": 0.005998397711664438, "learning_rate": 8.83022221559489e-05, "loss": -0.045487549155950546, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 18, "train_speed(iter/s)": 0.000432 }, @@ -298,7 +297,7 @@ "eval_reward_std": 0.10691346973180771, "eval_rewards/CosineReward": 0.03729327768087387, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1041.2321, + "eval_runtime": 1041.231, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 18 @@ -311,7 +310,7 @@ "kl": 0.1820068359375, "learning_rate": 8.636868207865244e-05, "loss": -0.03466903418302536, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.466796875, "reward": 0.04069916973821819, "reward_std": 0.11991005763411522, @@ -327,7 +326,7 @@ "kl": 0.19287109375, "learning_rate": 8.43120818934367e-05, "loss": -0.03502114117145538, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 20, "train_speed(iter/s)": 0.000424 }, @@ -339,14 +338,14 @@ "kl": 0.17626953125, "learning_rate": 8.213938048432697e-05, "loss": -0.008662773296236992, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.5625, "reward": 0.04996980866417289, "reward_std": 0.13849420100450516, "rewards/CosineReward": 0.049969930201768875, "rewards/RepetitionPenalty": -1.1864573679076784e-07, "step": 21, - "train_speed(iter/s)": 0.000407 + "train_speed(iter/s)": 0.000408 }, { "clip_ratio": 5.869188044016482e-05, @@ -355,7 +354,7 @@ "kl": 0.178955078125, "learning_rate": 7.985792958513931e-05, "loss": -0.008743642829358578, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 22, "train_speed(iter/s)": 0.000426 }, @@ -367,7 +366,7 @@ "kl": 0.1796875, "learning_rate": 7.74754489035403e-05, "loss": -0.03423420712351799, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.583984375, "reward": 0.034468831261619925, "reward_std": 0.11841745302081108, @@ -381,7 +380,7 @@ "grad_norm": 0.014131724834442139, "learning_rate": 7.500000000000001e-05, "loss": -0.03426633030176163, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 24, "train_speed(iter/s)": 0.000427 }, @@ -396,7 +395,7 @@ "eval_reward_std": 0.10456253588199615, "eval_rewards/CosineReward": 0.04339282959699631, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1045.0642, + "eval_runtime": 1045.0632, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 24 @@ -409,7 +408,7 @@ "kl": 0.1800537109375, "learning_rate": 7.243995901002312e-05, "loss": -0.02097315341234207, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.6171875, "reward": 0.03010205877944827, "reward_std": 0.10742511600255966, @@ -425,7 +424,7 @@ "kl": 0.18408203125, "learning_rate": 6.980398830195785e-05, "loss": -0.02103913575410843, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 26, "train_speed(iter/s)": 0.000421 }, @@ -437,7 +436,7 @@ "kl": 0.174560546875, "learning_rate": 6.710100716628344e-05, "loss": -0.03593946248292923, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.513671875, "reward": 0.04752760287374258, "reward_std": 0.14935147762298584, @@ -453,7 +452,7 @@ "kl": 0.182373046875, "learning_rate": 6.434016163555452e-05, "loss": -0.03595500811934471, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 28, "train_speed(iter/s)": 0.000422 }, @@ -465,7 +464,7 @@ "kl": 0.18701171875, "learning_rate": 6.153079353712201e-05, "loss": -0.031890563666820526, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "response_clip_ratio": 0.541015625, "reward": 0.04964290652424097, "reward_std": 0.1329497341066599, @@ -479,7 +478,7 @@ "grad_norm": 0.014435957185924053, "learning_rate": 5.868240888334653e-05, "loss": -0.032097991555929184, - "memory(GiB)": 187.02, + "memory(GiB)": 182.91, "step": 30, "train_speed(iter/s)": 0.000423 }, @@ -494,7 +493,7 @@ "eval_reward_std": 0.1401301473379135, "eval_rewards/CosineReward": 0.05227777361869812, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1085.6089, + "eval_runtime": 1085.6092, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 30 diff --git a/checkpoint-30/training_args.bin b/checkpoint-30/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-30/training_args.bin +++ b/checkpoint-30/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-30/zero_to_fp32.py b/checkpoint-30/zero_to_fp32.py new file mode 100644 index 0000000000000000000000000000000000000000..0e759146cadd92ddfefab3680146c2bd6a2b5c04 --- /dev/null +++ b/checkpoint-30/zero_to_fp32.py @@ -0,0 +1,760 @@ +#!/usr/bin/env python + +# Copyright (c) Microsoft Corporation. +# SPDX-License-Identifier: Apache-2.0 + +# DeepSpeed Team + +# This script extracts fp32 consolidated weights from a zero 1, 2 and 3 DeepSpeed checkpoints. It gets +# copied into the top level checkpoint dir, so the user can easily do the conversion at any point in +# the future. Once extracted, the weights don't require DeepSpeed and can be used in any +# application. +# +# example: +# python zero_to_fp32.py . output_dir/ +# or +# python zero_to_fp32.py . output_dir/ --safe_serialization + +import argparse +import torch +import glob +import math +import os +import re +import gc +import json +import numpy as np +from tqdm import tqdm +from collections import OrderedDict +from dataclasses import dataclass + +# while this script doesn't use deepspeed to recover data, since the checkpoints are pickled with +# DeepSpeed data structures it has to be available in the current python environment. +from deepspeed.utils import logger +from deepspeed.checkpoint.constants import (DS_VERSION, OPTIMIZER_STATE_DICT, SINGLE_PARTITION_OF_FP32_GROUPS, + FP32_FLAT_GROUPS, ZERO_STAGE, PARTITION_COUNT, PARAM_SHAPES, BUFFER_NAMES, + FROZEN_PARAM_SHAPES, FROZEN_PARAM_FRAGMENTS) + + +@dataclass +class zero_model_state: + buffers: dict() + param_shapes: dict() + shared_params: list + ds_version: int + frozen_param_shapes: dict() + frozen_param_fragments: dict() + + +debug = 0 + +# load to cpu +device = torch.device('cpu') + + +def atoi(text): + return int(text) if text.isdigit() else text + + +def natural_keys(text): + ''' + alist.sort(key=natural_keys) sorts in human order + http://nedbatchelder.com/blog/200712/human_sorting.html + (See Toothy's implementation in the comments) + ''' + return [atoi(c) for c in re.split(r'(\d+)', text)] + + +def get_model_state_file(checkpoint_dir, zero_stage): + if not os.path.isdir(checkpoint_dir): + raise FileNotFoundError(f"Directory '{checkpoint_dir}' doesn't exist") + + # there should be only one file + if zero_stage <= 2: + file = os.path.join(checkpoint_dir, "mp_rank_00_model_states.pt") + elif zero_stage == 3: + file = os.path.join(checkpoint_dir, "zero_pp_rank_0_mp_rank_00_model_states.pt") + + if not os.path.exists(file): + raise FileNotFoundError(f"can't find model states file at '{file}'") + + return file + + +def get_checkpoint_files(checkpoint_dir, glob_pattern): + # XXX: need to test that this simple glob rule works for multi-node setup too + ckpt_files = sorted(glob.glob(os.path.join(checkpoint_dir, glob_pattern)), key=natural_keys) + + if len(ckpt_files) == 0: + raise FileNotFoundError(f"can't find {glob_pattern} files in directory '{checkpoint_dir}'") + + return ckpt_files + + +def get_optim_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_optim_states.pt") + + +def get_model_state_files(checkpoint_dir): + return get_checkpoint_files(checkpoint_dir, "*_model_states.pt") + + +def parse_model_states(files): + zero_model_states = [] + for file in files: + state_dict = torch.load(file, map_location=device, weights_only=False) + + if BUFFER_NAMES not in state_dict: + raise ValueError(f"{file} is not a model state checkpoint") + buffer_names = state_dict[BUFFER_NAMES] + if debug: + print("Found buffers:", buffer_names) + + # recover just the buffers while restoring them to fp32 if they were saved in fp16 + buffers = {k: v.float() for k, v in state_dict["module"].items() if k in buffer_names} + param_shapes = state_dict[PARAM_SHAPES] + + # collect parameters that are included in param_shapes + param_names = [] + for s in param_shapes: + for name in s.keys(): + param_names.append(name) + + # update with frozen parameters + frozen_param_shapes = state_dict.get(FROZEN_PARAM_SHAPES, None) + if frozen_param_shapes is not None: + if debug: + print(f"Found frozen_param_shapes: {frozen_param_shapes}") + param_names += list(frozen_param_shapes.keys()) + + # handle shared params + shared_params = [[k, v] for k, v in state_dict["shared_params"].items()] + + ds_version = state_dict.get(DS_VERSION, None) + + frozen_param_fragments = state_dict.get(FROZEN_PARAM_FRAGMENTS, None) + + z_model_state = zero_model_state(buffers=buffers, + param_shapes=param_shapes, + shared_params=shared_params, + ds_version=ds_version, + frozen_param_shapes=frozen_param_shapes, + frozen_param_fragments=frozen_param_fragments) + zero_model_states.append(z_model_state) + + return zero_model_states + + +def parse_optim_states(files, ds_checkpoint_dir): + total_files = len(files) + state_dicts = [] + for f in tqdm(files, desc='Loading checkpoint shards'): + state_dict = torch.load(f, map_location=device, mmap=True, weights_only=False) + # immediately discard the potentially huge 2 optimizer states as we only care for fp32 master weights + # and also handle the case where it was already removed by another helper script + state_dict["optimizer_state_dict"].pop("optimizer_state_dict", None) + state_dicts.append(state_dict) + + if not ZERO_STAGE in state_dicts[0][OPTIMIZER_STATE_DICT]: + raise ValueError(f"{files[0]} is not a zero checkpoint") + zero_stage = state_dicts[0][OPTIMIZER_STATE_DICT][ZERO_STAGE] + world_size = state_dicts[0][OPTIMIZER_STATE_DICT][PARTITION_COUNT] + + # For ZeRO-2 each param group can have different partition_count as data parallelism for expert + # parameters can be different from data parallelism for non-expert parameters. So we can just + # use the max of the partition_count to get the dp world_size. + + if type(world_size) is list: + world_size = max(world_size) + + if world_size != total_files: + raise ValueError( + f"Expected {world_size} of '*_optim_states.pt' under '{ds_checkpoint_dir}' but found {total_files} files. " + "Possibly due to an overwrite of an old checkpoint, or a checkpoint didn't get saved by one or more processes." + ) + + # the groups are named differently in each stage + if zero_stage <= 2: + fp32_groups_key = SINGLE_PARTITION_OF_FP32_GROUPS + elif zero_stage == 3: + fp32_groups_key = FP32_FLAT_GROUPS + else: + raise ValueError(f"unknown zero stage {zero_stage}") + + fp32_flat_groups = [state_dicts[i][OPTIMIZER_STATE_DICT][fp32_groups_key] for i in range(len(state_dicts))] + return zero_stage, world_size, fp32_flat_groups + + +def _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters): + """ + Returns fp32 state_dict reconstructed from ds checkpoint + + Args: + - ``ds_checkpoint_dir``: path to the deepspeed checkpoint folder (where the optimizer files are) + + """ + print(f"Processing zero checkpoint '{ds_checkpoint_dir}'") + + optim_files = get_optim_files(ds_checkpoint_dir) + zero_stage, world_size, fp32_flat_groups = parse_optim_states(optim_files, ds_checkpoint_dir) + print(f"Detected checkpoint of type zero stage {zero_stage}, world_size: {world_size}") + + model_files = get_model_state_files(ds_checkpoint_dir) + + zero_model_states = parse_model_states(model_files) + print(f'Parsing checkpoint created by deepspeed=={zero_model_states[0].ds_version}') + + if zero_stage <= 2: + return _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + elif zero_stage == 3: + return _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters) + + +def _zero2_merge_frozen_params(state_dict, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + frozen_param_fragments = zero_model_states[0].frozen_param_fragments + + if debug: + num_elem = sum(s.numel() for s in frozen_param_shapes.values()) + print(f'rank 0: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in frozen_param_fragments.values()]) + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + state_dict[name] = frozen_param_fragments[name] + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +def _has_callable(obj, fn): + attr = getattr(obj, fn, None) + return callable(attr) + + +def _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + + # Reconstruction protocol: + # + # XXX: document this + + if debug: + for i in range(world_size): + for j in range(len(fp32_flat_groups[0])): + print(f"{FP32_FLAT_GROUPS}[{i}][{j}].shape={fp32_flat_groups[i][j].shape}") + + # XXX: memory usage doubles here (zero2) + num_param_groups = len(fp32_flat_groups[0]) + merged_single_partition_of_fp32_groups = [] + for i in range(num_param_groups): + merged_partitions = [sd[i] for sd in fp32_flat_groups] + full_single_fp32_vector = torch.cat(merged_partitions, 0) + merged_single_partition_of_fp32_groups.append(full_single_fp32_vector) + avail_numel = sum( + [full_single_fp32_vector.numel() for full_single_fp32_vector in merged_single_partition_of_fp32_groups]) + + if debug: + wanted_params = sum([len(shapes) for shapes in param_shapes]) + wanted_numel = sum([sum(shape.numel() for shape in shapes.values()) for shapes in param_shapes]) + # not asserting if there is a mismatch due to possible padding + print(f"Have {avail_numel} numels to process.") + print(f"Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + total_numel = 0 + total_params = 0 + for shapes, full_single_fp32_vector in zip(param_shapes, merged_single_partition_of_fp32_groups): + offset = 0 + avail_numel = full_single_fp32_vector.numel() + for name, shape in shapes.items(): + + unpartitioned_numel = shape.numel() if _has_callable(shape, 'numel') else math.prod(shape) + total_numel += unpartitioned_numel + total_params += 1 + + if debug: + print(f"{name} full shape: {shape} unpartitioned numel {unpartitioned_numel} ") + state_dict[name] = full_single_fp32_vector.narrow(0, offset, unpartitioned_numel).view(shape) + offset += unpartitioned_numel + + # Z2 started to align to 2*world_size to improve nccl performance. Therefore both offset and + # avail_numel can differ by anywhere between 0..2*world_size. Due to two unrelated complex + # paddings performed in the code it's almost impossible to predict the exact numbers w/o the + # live optimizer object, so we are checking that the numbers are within the right range + align_to = 2 * world_size + + def zero2_align(x): + return align_to * math.ceil(x / align_to) + + if debug: + print(f"original offset={offset}, avail_numel={avail_numel}") + + offset = zero2_align(offset) + avail_numel = zero2_align(avail_numel) + + if debug: + print(f"aligned offset={offset}, avail_numel={avail_numel}") + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero2_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero2_merge_frozen_params(state_dict, zero_model_states) + + _zero2_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def zero3_partitioned_param_info(unpartitioned_numel, world_size): + remainder = unpartitioned_numel % world_size + padding_numel = (world_size - remainder) if remainder else 0 + partitioned_numel = math.ceil(unpartitioned_numel / world_size) + return partitioned_numel, padding_numel + + +def _zero3_merge_frozen_params(state_dict, world_size, zero_model_states): + if zero_model_states[0].frozen_param_shapes is None or len(zero_model_states[0].frozen_param_shapes) == 0: + return + + if debug: + for i in range(world_size): + num_elem = sum(s.numel() for s in zero_model_states[i].frozen_param_fragments.values()) + print(f'rank {i}: {FROZEN_PARAM_SHAPES}.numel = {num_elem}') + + frozen_param_shapes = zero_model_states[0].frozen_param_shapes + wanted_params = len(frozen_param_shapes) + wanted_numel = sum(s.numel() for s in frozen_param_shapes.values()) + avail_numel = sum([p.numel() for p in zero_model_states[0].frozen_param_fragments.values()]) * world_size + print(f'Frozen params: Have {avail_numel} numels to process.') + print(f'Frozen params: Need {wanted_numel} numels in {wanted_params} params') + + total_params = 0 + total_numel = 0 + for name, shape in zero_model_states[0].frozen_param_shapes.items(): + total_params += 1 + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + + param_frags = tuple(model_state.frozen_param_fragments[name] for model_state in zero_model_states) + state_dict[name] = torch.cat(param_frags, 0).narrow(0, 0, unpartitioned_numel).view(shape) + + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Frozen params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + print(f"Reconstructed Frozen fp32 state dict with {total_params} params {total_numel} elements") + + +class GatheredTensor: + """ + A pseudo tensor that collects partitioned weights. + It is more memory efficient when there are multiple groups. + """ + + def __init__(self, flat_groups, flat_groups_offset, offset, partitioned_numel, shape): + self.flat_groups = flat_groups + self.flat_groups_offset = flat_groups_offset + self.offset = offset + self.partitioned_numel = partitioned_numel + self.shape = shape + self.dtype = self.flat_groups[0][0].dtype + + def contiguous(self): + """ + Merge partitioned weights from flat_groups into a single tensor. + """ + end_idx = self.offset + self.partitioned_numel + world_size = len(self.flat_groups) + pad_flat_param_chunks = [] + + for rank_i in range(world_size): + # for each rank, we need to collect weights from related group/groups + flat_groups_at_rank_i = self.flat_groups[rank_i] + start_group_id = None + end_group_id = None + for group_id in range(len(self.flat_groups_offset)): + if self.flat_groups_offset[group_id] <= self.offset < self.flat_groups_offset[group_id + 1]: + start_group_id = group_id + if self.flat_groups_offset[group_id] < end_idx <= self.flat_groups_offset[group_id + 1]: + end_group_id = group_id + break + # collect weights from related group/groups + for group_id in range(start_group_id, end_group_id + 1): + flat_tensor = flat_groups_at_rank_i[group_id] + start_offset = self.offset - self.flat_groups_offset[group_id] + end_offset = min(end_idx, self.flat_groups_offset[group_id + 1]) - self.flat_groups_offset[group_id] + pad_flat_param_chunks.append(flat_tensor[start_offset:end_offset]) + + # collect weights from all ranks + pad_flat_param = torch.cat(pad_flat_param_chunks, dim=0) + param = pad_flat_param[:self.shape.numel()].view(self.shape).contiguous() + return param + + +def _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states): + param_shapes = zero_model_states[0].param_shapes + avail_numel = sum([flat_group.numel() for flat_group in fp32_flat_groups[0]]) * world_size + + # Reconstruction protocol: For zero3 we need to zip the partitions together at boundary of each + # param, re-consolidating each param, while dealing with padding if any + + # merge list of dicts, preserving order + param_shapes = {k: v for d in param_shapes for k, v in d.items()} + + if debug: + for i in range(world_size): + print(f"{FP32_FLAT_GROUPS}[{i}].shape={fp32_flat_groups[i].shape}") + + wanted_params = len(param_shapes) + wanted_numel = sum(shape.numel() for shape in param_shapes.values()) + # not asserting if there is a mismatch due to possible padding + avail_numel = fp32_flat_groups[0].numel() * world_size + print(f"Trainable params: Have {avail_numel} numels to process.") + print(f"Trainable params: Need {wanted_numel} numels in {wanted_params} params.") + + # params + # XXX: for huge models that can't fit into the host's RAM we will have to recode this to support + # out-of-core computing solution + offset = 0 + total_numel = 0 + total_params = 0 + flat_groups_offset = [0] + list(np.cumsum([flat_tensor.numel() for flat_tensor in fp32_flat_groups[0]])) + for name, shape in tqdm(param_shapes.items(), desc='Gathering sharded weights'): + unpartitioned_numel = shape.numel() + total_numel += unpartitioned_numel + total_params += 1 + partitioned_numel, partitioned_padding_numel = zero3_partitioned_param_info(unpartitioned_numel, world_size) + + if debug: + print( + f"Trainable params: {total_params} {name} full shape: {shape} partition0 numel={partitioned_numel} partitioned_padding_numel={partitioned_padding_numel}" + ) + + # memory efficient tensor + tensor = GatheredTensor(fp32_flat_groups, flat_groups_offset, offset, partitioned_numel, shape) + state_dict[name] = tensor + offset += partitioned_numel + + offset *= world_size + + # Sanity check + if offset != avail_numel: + raise ValueError(f"consumed {offset} numels out of {avail_numel} - something is wrong") + + print(f"Reconstructed Trainable fp32 state dict with {total_params} params {total_numel} elements") + + +def _get_fp32_state_dict_from_zero3_checkpoint(world_size, fp32_flat_groups, zero_model_states, + exclude_frozen_parameters): + state_dict = OrderedDict() + + # buffers + buffers = zero_model_states[0].buffers + state_dict.update(buffers) + if debug: + print(f"added {len(buffers)} buffers") + + if not exclude_frozen_parameters: + _zero3_merge_frozen_params(state_dict, world_size, zero_model_states) + + _zero3_merge_trainable_params(state_dict, world_size, fp32_flat_groups, zero_model_states) + + # recover shared parameters + for pair in zero_model_states[0].shared_params: + if pair[1] in state_dict: + state_dict[pair[0]] = state_dict[pair[1]] + + return state_dict + + +def to_torch_tensor(state_dict, return_empty_tensor=False): + """ + Convert state_dict of GatheredTensor to torch tensor + """ + torch_state_dict = {} + converted_tensors = {} + for name, tensor in state_dict.items(): + tensor_id = id(tensor) + if tensor_id in converted_tensors: # shared tensors + shared_tensor = torch_state_dict[converted_tensors[tensor_id]] + torch_state_dict[name] = shared_tensor + else: + converted_tensors[tensor_id] = name + if return_empty_tensor: + torch_state_dict[name] = torch.empty(tensor.shape, dtype=tensor.dtype) + else: + torch_state_dict[name] = tensor.contiguous() + return torch_state_dict + + +def get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, + tag=None, + exclude_frozen_parameters=False, + lazy_mode=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated state_dict that can be loaded with + ``load_state_dict()`` and used for training without DeepSpeed or shared with others, for example + via a model hub. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in 'latest' file. e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + - ``lazy_mode``: get state_dict in lazy mode. It returns a dict of pesduo tensor instead of torch tensor, which is more memory efficient. + Convert the pesduo tensor to torch tensor by ``.contiguous()`` + + Returns: + - pytorch ``state_dict`` + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + # do the training and checkpoint saving + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir) # already on cpu + model = model.cpu() # move to cpu + model.load_state_dict(state_dict) + # submit to model hub or save the model to share with others + + In this example the ``model`` will no longer be usable in the deepspeed context of the same + application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + If you want it all done for you, use ``load_state_dict_from_zero_checkpoint`` instead. + + Note: the above usage may not work if your application doesn't have sufficient free CPU memory. + You may need to use the offline approach using the ``zero_to_fp32.py`` script that is saved with + the checkpoint. Or you can load state_dict in lazy mode :: + + from deepspeed.utils.zero_to_fp32 import get_fp32_state_dict_from_zero_checkpoint + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, lazy_mode=True) # not on cpu + for name, lazy_tensor in state_dict.item(): + tensor = lazy_tensor.contiguous() # to cpu + print(name, tensor) + # del tensor to release memory if it no longer in use + """ + if tag is None: + latest_path = os.path.join(checkpoint_dir, 'latest') + if os.path.isfile(latest_path): + with open(latest_path, 'r') as fd: + tag = fd.read().strip() + else: + raise ValueError(f"Unable to find 'latest' file at {latest_path}") + + ds_checkpoint_dir = os.path.join(checkpoint_dir, tag) + + if not os.path.isdir(ds_checkpoint_dir): + raise FileNotFoundError(f"Directory '{ds_checkpoint_dir}' doesn't exist") + + state_dict = _get_fp32_state_dict_from_zero_checkpoint(ds_checkpoint_dir, exclude_frozen_parameters) + if lazy_mode: + return state_dict + else: + return to_torch_tensor(state_dict) + + +def convert_zero_checkpoint_to_fp32_state_dict(checkpoint_dir, + output_dir, + max_shard_size="5GB", + safe_serialization=False, + tag=None, + exclude_frozen_parameters=False): + """ + Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` file that can be + loaded with ``torch.load(file)`` + ``load_state_dict()`` and used for training without DeepSpeed. + + Args: + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``output_dir``: directory to the pytorch fp32 state_dict output files + - ``max_shard_size``: the maximum size for a checkpoint before being sharded, default value is 5GB + - ``safe_serialization``: whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`). + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + - ``exclude_frozen_parameters``: exclude frozen parameters + """ + + # Dependency pre-check + if safe_serialization: + try: + from safetensors.torch import save_file + except ImportError: + print('If you want to use `safe_serialization`, please `pip install safetensors`') + raise + if max_shard_size is not None: + try: + from huggingface_hub import split_torch_state_dict_into_shards + except ImportError: + print('If you want to use `max_shard_size`, please `pip install huggingface_hub`') + raise + + # Convert zero checkpoint to state_dict + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, + tag, + exclude_frozen_parameters, + lazy_mode=True) + + # Shard the model if it is too big. + weights_name = "model.safetensors" if safe_serialization else "pytorch_model.bin" + if max_shard_size is not None: + filename_pattern = weights_name.replace(".bin", "{suffix}.bin").replace(".safetensors", "{suffix}.safetensors") + # an memory-efficient approach for sharding + empty_state_dict = to_torch_tensor(state_dict, return_empty_tensor=True) + state_dict_split = split_torch_state_dict_into_shards(empty_state_dict, + filename_pattern=filename_pattern, + max_shard_size=max_shard_size) + else: + from collections import namedtuple + StateDictSplit = namedtuple("StateDictSplit", ["is_sharded", "filename_to_tensors"]) + state_dict_split = StateDictSplit(is_sharded=False, + filename_to_tensors={weights_name: list(state_dict.keys())}) + + # Save the model by shard + os.makedirs(output_dir, exist_ok=True) + filename_to_tensors = state_dict_split.filename_to_tensors.items() + for shard_file, tensors in tqdm(filename_to_tensors, desc="Saving checkpoint shards"): + shard_state_dict = {tensor_name: state_dict[tensor_name] for tensor_name in tensors} + shard_state_dict = to_torch_tensor(shard_state_dict) + output_path = os.path.join(output_dir, shard_file) + if safe_serialization: + save_file(shard_state_dict, output_path, metadata={"format": "pt"}) + else: + torch.save(shard_state_dict, output_path) + # release the memory of current shard + for tensor_name in list(shard_state_dict.keys()): + del state_dict[tensor_name] + del shard_state_dict[tensor_name] + del shard_state_dict + gc.collect() + + # Save index if sharded + if state_dict_split.is_sharded: + index = { + "metadata": state_dict_split.metadata, + "weight_map": state_dict_split.tensor_to_filename, + } + save_index_file = "model.safetensors.index.json" if safe_serialization else "pytorch_model.bin.index.json" + save_index_file = os.path.join(output_dir, save_index_file) + with open(save_index_file, "w", encoding="utf-8") as f: + content = json.dumps(index, indent=2, sort_keys=True) + "\n" + f.write(content) + + +def load_state_dict_from_zero_checkpoint(model, checkpoint_dir, tag=None): + """ + 1. Put the provided model to cpu + 2. Convert ZeRO 2 or 3 checkpoint into a single fp32 consolidated ``state_dict`` + 3. Load it into the provided model + + Args: + - ``model``: the model object to update + - ``checkpoint_dir``: path to the desired checkpoint folder. (one that contains the tag-folder, like ``global_step14``) + - ``tag``: checkpoint tag used as a unique identifier for checkpoint. If not provided will attempt to load tag in the file named ``latest`` in the checkpoint folder, e.g., ``global_step14`` + + Returns: + - ``model`: modified model + + Make sure you have plenty of CPU memory available before you call this function. If you don't + have enough use the ``zero_to_fp32.py`` utility to do the conversion. You will find it + conveniently placed for you in the checkpoint folder. + + A typical usage might be :: + + from deepspeed.utils.zero_to_fp32 import load_state_dict_from_zero_checkpoint + model = load_state_dict_from_zero_checkpoint(trainer.model, checkpoint_dir) + # submit to model hub or save the model to share with others + + Note, that once this was run, the ``model`` will no longer be usable in the deepspeed context + of the same application. i.e. you will need to re-initialize the deepspeed engine, since + ``model.load_state_dict(state_dict)`` will remove all the deepspeed magic from it. + + """ + logger.info(f"Extracting fp32 weights") + state_dict = get_fp32_state_dict_from_zero_checkpoint(checkpoint_dir, tag) + + logger.info(f"Overwriting model with fp32 weights") + model = model.cpu() + model.load_state_dict(state_dict, strict=False) + + return model + + +if __name__ == "__main__": + parser = argparse.ArgumentParser() + parser.add_argument("checkpoint_dir", + type=str, + help="path to the desired checkpoint folder, e.g., path/checkpoint-12") + parser.add_argument("output_dir", + type=str, + help="directory to the pytorch fp32 state_dict output files" + "(e.g. path/checkpoint-12-output/)") + parser.add_argument( + "--max_shard_size", + type=str, + default="5GB", + help="The maximum size for a checkpoint before being sharded. Checkpoints shard will then be each of size" + "lower than this size. If expressed as a string, needs to be digits followed by a unit (like `5MB`" + "We default it to 5GB in order for models to be able to run easily on free-tier google colab instances" + "without CPU OOM issues.") + parser.add_argument( + "--safe_serialization", + default=False, + action='store_true', + help="Whether to save the model using `safetensors` or the traditional PyTorch way (that uses `pickle`).") + parser.add_argument("-t", + "--tag", + type=str, + default=None, + help="checkpoint tag used as a unique identifier for checkpoint. e.g., global_step1") + parser.add_argument("--exclude_frozen_parameters", action='store_true', help="exclude frozen parameters") + parser.add_argument("-d", "--debug", action='store_true', help="enable debug") + args = parser.parse_args() + + debug = args.debug + + convert_zero_checkpoint_to_fp32_state_dict(args.checkpoint_dir, + args.output_dir, + max_shard_size=args.max_shard_size, + safe_serialization=args.safe_serialization, + tag=args.tag, + exclude_frozen_parameters=args.exclude_frozen_parameters) diff --git a/checkpoint-4/adapter_config.json b/checkpoint-4/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-4/adapter_config.json +++ b/checkpoint-4/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-4/trainer_state.json b/checkpoint-4/trainer_state.json index b95d617400d7530f7b064a18f7bc7cdc945a2aef..f595f5961ff8df694c28b2d941589615801e0294 100644 --- a/checkpoint-4/trainer_state.json +++ b/checkpoint-4/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.8421052631578947, @@ -7,7 +6,7 @@ "global_step": 4, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 } diff --git a/checkpoint-4/training_args.bin b/checkpoint-4/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-4/training_args.bin +++ b/checkpoint-4/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-6/adapter_config.json b/checkpoint-6/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-6/adapter_config.json +++ b/checkpoint-6/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-6/trainer_state.json b/checkpoint-6/trainer_state.json index d4540222f113714a1953b0e161b891639c52298f..3667b41328b83c70faf4e00796621b5ccfe72baf 100644 --- a/checkpoint-6/trainer_state.json +++ b/checkpoint-6/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 6, "best_metric": 0.012996690347790718, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-6", "epoch": 1.4210526315789473, @@ -7,7 +6,7 @@ "global_step": 6, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 diff --git a/checkpoint-6/training_args.bin b/checkpoint-6/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-6/training_args.bin +++ b/checkpoint-6/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/checkpoint-8/adapter_config.json b/checkpoint-8/adapter_config.json index 9a6753254d3cb15865ffb290b9553d85b57dac9a..3f8d06e83be137d4af154849ed1686625c42e280 100644 --- a/checkpoint-8/adapter_config.json +++ b/checkpoint-8/adapter_config.json @@ -24,10 +24,10 @@ "revision": null, "target_modules": [ "v_proj", - "up_proj", - "k_proj", "gate_proj", + "k_proj", "o_proj", + "up_proj", "q_proj", "down_proj" ], diff --git a/checkpoint-8/trainer_state.json b/checkpoint-8/trainer_state.json index 3a6b398642c37165b24d317105212cefd3a27a22..1878466a3ed5546ada8088e46201029d0f48bef5 100644 --- a/checkpoint-8/trainer_state.json +++ b/checkpoint-8/trainer_state.json @@ -1,5 +1,4 @@ { - "best_global_step": 6, "best_metric": 0.012996690347790718, "best_model_checkpoint": "/mnt/nvme5n1p1/trained_grpo_distill_14b_rl_70_s3/v3-20250330-200345/checkpoint-6", "epoch": 1.8421052631578947, @@ -7,7 +6,7 @@ "global_step": 8, "is_hyper_param_search": false, "is_local_process_zero": true, - "is_world_process_zero": false, + "is_world_process_zero": true, "log_history": [ { "clip_ratio": 0.0, @@ -17,14 +16,14 @@ "kl": 0.0, "learning_rate": 1.6666666666666667e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.11328125, "reward": -0.002658387296833098, "reward_std": 0.06134121119976044, "rewards/CosineReward": -0.0026579967816360295, "rewards/RepetitionPenalty": -3.8975886695880035e-07, "step": 1, - "train_speed(iter/s)": 0.000241 + "train_speed(iter/s)": 0.000242 }, { "clip_ratio": 0.0, @@ -33,9 +32,9 @@ "kl": 0.0, "learning_rate": 3.3333333333333335e-05, "loss": -0.11016345024108887, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 2, - "train_speed(iter/s)": 0.000466 + "train_speed(iter/s)": 0.000467 }, { "clip_ratio": 1.3441811461234465e-05, @@ -45,7 +44,7 @@ "kl": 9.50181856751442e-07, "learning_rate": 5e-05, "loss": -0.06604708731174469, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.13671875, "reward": 0.0006296975770965219, "reward_std": 0.07172460854053497, @@ -61,7 +60,7 @@ "kl": 1.1101365089416504e-05, "learning_rate": 6.666666666666667e-05, "loss": -0.06727766245603561, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 4, "train_speed(iter/s)": 0.000458 }, @@ -73,7 +72,7 @@ "kl": 0.00017762184143066406, "learning_rate": 8.333333333333334e-05, "loss": -0.09315311908721924, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.119140625, "reward": -0.005135859013535082, "reward_std": 0.07994875870645046, @@ -87,9 +86,9 @@ "grad_norm": 0.18263348937034607, "learning_rate": 0.0001, "loss": -0.1041698157787323, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 6, - "train_speed(iter/s)": 0.000458 + "train_speed(iter/s)": 0.000459 }, { "epoch": 1.4210526315789473, @@ -102,7 +101,7 @@ "eval_reward_std": 0.08769983053207397, "eval_rewards/CosineReward": 0.012996694073081017, "eval_rewards/RepetitionPenalty": 0.0, - "eval_runtime": 1030.1126, + "eval_runtime": 1030.1127, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "step": 6 @@ -115,7 +114,7 @@ "kl": 0.017406463623046875, "learning_rate": 9.991540791356342e-05, "loss": -0.051375165581703186, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "response_clip_ratio": 0.1484375, "reward": 0.004909618757665157, "reward_std": 0.08167182095348835, @@ -131,7 +130,7 @@ "kl": 0.089599609375, "learning_rate": 9.966191788709716e-05, "loss": -0.05105742812156677, - "memory(GiB)": 176.98, + "memory(GiB)": 182.91, "step": 8, "train_speed(iter/s)": 0.000433 } diff --git a/checkpoint-8/training_args.bin b/checkpoint-8/training_args.bin index db6dc02fc187938d09e0a6626a08092ad8f42c54..a5c75ad76398d7f403ccc1a74c463d2dab6465ca 100644 --- a/checkpoint-8/training_args.bin +++ b/checkpoint-8/training_args.bin @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:09cdf21dfd9faa218b7fd99e3f3dc0ef681c4e3fd3b905e7348f5467b0198044 +oid sha256:1207fcb9d91c7deb13a80104f3ca89016b4cff3ef13ebd136ee6320d5a9888bb size 9809 diff --git a/completions.jsonl b/completions.jsonl index 97b3e9dc3878a77a5dff83d2f2c280879a247d35..999249887b4ecd2d23dfd4e6feb0d87167ff4fd1 100644 --- a/completions.jsonl +++ b/completions.jsonl @@ -1,3 +1,3 @@ version https://git-lfs.github.com/spec/v1 -oid sha256:d0cac5ba83180c1a0721b5e0604d154eda7f5a1aee67b1ced24357e9e6c7e175 -size 272303397 +oid sha256:511b8afad392a7fca01151de5a1a815f9cd4bb830e428d6470ebf64206fc4d47 +size 276681086 diff --git a/logging.jsonl b/logging.jsonl index f5c2129421a0917ea55a5da41e0ceefbdcb02c82..c368b67f1a71e64806cfcd23006f3ecf64d3b016 100644 --- a/logging.jsonl +++ b/logging.jsonl @@ -32,3 +32,4 @@ {"loss": -0.03595501, "grad_norm": 0.00527766, "learning_rate": 6.434e-05, "memory(GiB)": 182.91, "train_speed(iter/s)": 0.000422, "kl": 0.18237305, "clip_ratio": 5.954e-05, "epoch": 6.84210526, "global_step/max_steps": "28/60", "percentage": "46.67%", "elapsed_time": "18h 25m 42s", "remaining_time": "21h 3m 40s"} {"loss": -0.03189056, "grad_norm": 0.0168444, "learning_rate": 6.153e-05, "memory(GiB)": 182.91, "train_speed(iter/s)": 0.00041, "completion_length": 10427.296875, "response_clip_ratio": 0.54101562, "rewards/CosineReward": 0.04964366, "rewards/RepetitionPenalty": -7.5e-07, "reward": 0.04964291, "reward_std": 0.13294973, "kl": 0.18701172, "clip_ratio": 5.28e-05, "epoch": 7.21052632, "global_step/max_steps": "29/60", "percentage": "48.33%", "elapsed_time": "19h 37m 59s", "remaining_time": "20h 59m 14s"} {"loss": -0.03209799, "grad_norm": 0.01443596, "learning_rate": 5.868e-05, "memory(GiB)": 182.91, "train_speed(iter/s)": 0.000423, "epoch": 7.42105263, "global_step/max_steps": "30/60", "percentage": "50.00%", "elapsed_time": "19h 40m 53s", "remaining_time": "19h 40m 53s"} +{"eval_loss": -0.09817081, "eval_completion_length": 12289.0, "eval_response_clip_ratio": 1.0, "eval_rewards/CosineReward": 0.05227777, "eval_rewards/RepetitionPenalty": 0.0, "eval_reward": 0.05227778, "eval_reward_std": 0.14013015, "eval_kl": 0.19726562, "eval_clip_ratio": 4.542e-05, "eval_runtime": 1085.6092, "eval_samples_per_second": 0.001, "eval_steps_per_second": 0.001, "epoch": 7.42105263, "global_step/max_steps": "30/60", "percentage": "50.00%", "elapsed_time": "19h 58m 59s", "remaining_time": "19h 58m 59s"}