Training in progress, epoch 0, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5c7d696805c33619b84f2fedbcdd5a68607d7306abf5dbf4533f634e4982f5c8
+oid sha256:6a9c3e8764ac0e745f6f2f52febb7d10692a33c3745620483b24136f9b8a908e
 size 1172343536

last-checkpoint/global_step250/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8179a1613896aa79455e50e4aa8d77bd60a506db72324d864e9cdca54cc60626
+size 883824229

last-checkpoint/global_step250/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9fc5a87dbe8a7cce6e57a9b3ace28fd4f4f5928893f68f6debd75d62a4729109
+size 883824293

last-checkpoint/global_step250/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f48a186df598e4b9071facf76f95274858f30d911693188fcb82ef9ed3e9ecf
+size 883824293

last-checkpoint/global_step250/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:a0cb4a60e6514af9a7d74cbcd436350c03f46da29e427d75838e9697d64139ec
+size 883824293

last-checkpoint/global_step250/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4dc344d017fe7374aa6e0dcc3680a9c16ccf9c24de138199ee5b5bb69d19fdd9
+size 1172522073

last-checkpoint/latest

@@ -1 +1 @@
-global_step200
+global_step250

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:81d5f83aeb4b3f559bd28377336d47659b320e7f6ef2e5a723d284716278a151
+oid sha256:9892a41432d4e48929c5376b8225d5373bb9afe9f14665fbcb3934dc45cb330a
 size 15429

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2626437dcb133ffcf003ac89603f8cce07459b93a98d760cd9419e0d6a994067
+oid sha256:b6f3949e2b25ad05905da16bc2d58aeadc9a9f1db0b0e97a04ceb63467acff1d
 size 15429

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ae777e24d50cb7159634e1245f0697ba0fc64d5b26d535f2c80e411371a90b1c
+oid sha256:5c3812c5e7f13c9d32f97d83cd625428061b1c9486e869caffd453d83ac07ceb
 size 15429

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:afc5a67564eebcfc961e8f1406a7418cc73497c2935a39af0232ef59f8153a6a
+oid sha256:6eda756b33768639c1e1932ae00c5d43269de5705c09162bc09acdf71d3533c0
 size 15429

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:331d42834d40f6ab9fe0c79a5f67902cd952ec456fbe0c511bb4b4fc0a7cf69b
+oid sha256:610f32c595d4257f7e544a8d61dbdf0e85b19520c0d69daf98d6e00fbe896bb3
 size 1401

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
   "best_global_step": null,
-  "best_metric": 0.06463618576526642,
+  "best_metric": 0.05831117928028107,
   "best_model_checkpoint": null,
-  "epoch": 0.1737242128121607,
+  "epoch": 0.21715526601520088,
   "eval_steps": 50,
-  "global_step": 200,
+  "global_step": 250,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -672,6 +672,172 @@
       "eval_samples_per_second": 45.248,
       "eval_steps_per_second": 2.836,
       "step": 200
+    },
+    {
+      "epoch": 0.1780673181324647,
+      "grad_norm": 2.3196284770965576,
+      "learning_rate": 2.5983748711161684e-05,
+      "logits/chosen": -0.4415039122104645,
+      "logits/rejected": -0.7743164300918579,
+      "logps/chosen": -104.36250305175781,
+      "logps/rejected": -187.5,
+      "loss": 0.0463,
+      "rewards/accuracies": 0.981249988079071,
+      "rewards/chosen": -3.2431640625,
+      "rewards/margins": 11.009374618530273,
+      "rewards/rejected": -14.260937690734863,
+      "step": 205
+    },
+    {
+      "epoch": 0.18241042345276873,
+      "grad_norm": 1.842193365097046,
+      "learning_rate": 2.598305199430297e-05,
+      "logits/chosen": -0.49321287870407104,
+      "logits/rejected": -0.8467773199081421,
+      "logps/chosen": -104.80000305175781,
+      "logps/rejected": -197.0,
+      "loss": 0.0531,
+      "rewards/accuracies": 0.981249988079071,
+      "rewards/chosen": -3.1109375953674316,
+      "rewards/margins": 12.276562690734863,
+      "rewards/rejected": -15.381250381469727,
+      "step": 210
+    },
+    {
+      "epoch": 0.18675352877307275,
+      "grad_norm": 1.857373595237732,
+      "learning_rate": 2.5982322581077525e-05,
+      "logits/chosen": -0.534863293170929,
+      "logits/rejected": -0.8707031011581421,
+      "logps/chosen": -104.2249984741211,
+      "logps/rejected": -199.3249969482422,
+      "loss": 0.0556,
+      "rewards/accuracies": 0.971875011920929,
+      "rewards/chosen": -3.1683592796325684,
+      "rewards/margins": 12.2265625,
+      "rewards/rejected": -15.395312309265137,
+      "step": 215
+    },
+    {
+      "epoch": 0.19109663409337677,
+      "grad_norm": 1.5108240842819214,
+      "learning_rate": 2.598156047332165e-05,
+      "logits/chosen": -0.49296873807907104,
+      "logits/rejected": -0.8592773675918579,
+      "logps/chosen": -99.38749694824219,
+      "logps/rejected": -190.1750030517578,
+      "loss": 0.0509,
+      "rewards/accuracies": 0.979687511920929,
+      "rewards/chosen": -2.543261766433716,
+      "rewards/margins": 11.925000190734863,
+      "rewards/rejected": -14.473437309265137,
+      "step": 220
+    },
+    {
+      "epoch": 0.19543973941368079,
+      "grad_norm": 2.208911418914795,
+      "learning_rate": 2.5980765672953946e-05,
+      "logits/chosen": -0.48857420682907104,
+      "logits/rejected": -0.869824230670929,
+      "logps/chosen": -99.2249984741211,
+      "logps/rejected": -189.39999389648438,
+      "loss": 0.0698,
+      "rewards/accuracies": 0.9750000238418579,
+      "rewards/chosen": -2.405956983566284,
+      "rewards/margins": 12.265625,
+      "rewards/rejected": -14.670312881469727,
+      "step": 225
+    },
+    {
+      "epoch": 0.1997828447339848,
+      "grad_norm": 1.6780784130096436,
+      "learning_rate": 2.5979938181975324e-05,
+      "logits/chosen": -0.537060558795929,
+      "logits/rejected": -0.892773449420929,
+      "logps/chosen": -100.7125015258789,
+      "logps/rejected": -199.625,
+      "loss": 0.0294,
+      "rewards/accuracies": 0.9828125238418579,
+      "rewards/chosen": -2.6998047828674316,
+      "rewards/margins": 13.115625381469727,
+      "rewards/rejected": -15.807812690734863,
+      "step": 230
+    },
+    {
+      "epoch": 0.20412595005428882,
+      "grad_norm": 1.3097057342529297,
+      "learning_rate": 2.5979078002468992e-05,
+      "logits/chosen": -0.5240722894668579,
+      "logits/rejected": -0.8931640386581421,
+      "logps/chosen": -102.375,
+      "logps/rejected": -197.4250030517578,
+      "loss": 0.0603,
+      "rewards/accuracies": 0.979687511920929,
+      "rewards/chosen": -2.710986375808716,
+      "rewards/margins": 12.909375190734863,
+      "rewards/rejected": -15.634374618530273,
+      "step": 235
+    },
+    {
+      "epoch": 0.20846905537459284,
+      "grad_norm": 2.499241828918457,
+      "learning_rate": 2.597818513660045e-05,
+      "logits/chosen": -0.4613281190395355,
+      "logits/rejected": -0.834277331829071,
+      "logps/chosen": -98.8499984741211,
+      "logps/rejected": -185.8249969482422,
+      "loss": 0.0777,
+      "rewards/accuracies": 0.9750000238418579,
+      "rewards/chosen": -2.1981444358825684,
+      "rewards/margins": 11.0859375,
+      "rewards/rejected": -13.290624618530273,
+      "step": 240
+    },
+    {
+      "epoch": 0.21281216069489686,
+      "grad_norm": 1.4439401626586914,
+      "learning_rate": 2.5977259586617493e-05,
+      "logits/chosen": -0.35028076171875,
+      "logits/rejected": -0.7183593511581421,
+      "logps/chosen": -95.6500015258789,
+      "logps/rejected": -176.0,
+      "loss": 0.0725,
+      "rewards/accuracies": 0.971875011920929,
+      "rewards/chosen": -2.0106444358825684,
+      "rewards/margins": 9.563281059265137,
+      "rewards/rejected": -11.579687118530273,
+      "step": 245
+    },
+    {
+      "epoch": 0.21715526601520088,
+      "grad_norm": 0.8243131041526794,
+      "learning_rate": 2.5976301354850176e-05,
+      "logits/chosen": -0.345947265625,
+      "logits/rejected": -0.679638683795929,
+      "logps/chosen": -98.36250305175781,
+      "logps/rejected": -182.125,
+      "loss": 0.0622,
+      "rewards/accuracies": 0.981249988079071,
+      "rewards/chosen": -2.5238280296325684,
+      "rewards/margins": 10.37109375,
+      "rewards/rejected": -12.893750190734863,
+      "step": 250
+    },
+    {
+      "epoch": 0.21715526601520088,
+      "eval_logits/chosen": -0.42717909812927246,
+      "eval_logits/rejected": -0.7368005514144897,
+      "eval_logps/chosen": -103.40794372558594,
+      "eval_logps/rejected": -193.13717651367188,
+      "eval_loss": 0.05831117928028107,
+      "eval_rewards/accuracies": 0.9769855737686157,
+      "eval_rewards/chosen": -3.0707638263702393,
+      "eval_rewards/margins": 11.707694053649902,
+      "eval_rewards/rejected": -14.779783248901367,
+      "eval_runtime": 97.7681,
+      "eval_samples_per_second": 45.209,
+      "eval_steps_per_second": 2.833,
+      "step": 250
     }
   ],
   "logging_steps": 5,