Training in progress, epoch 3, checkpoint

last-checkpoint/global_step189/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:451138fe67af60cef7cb1b4ebcd9fd2c7e03fd15c0d820244d76b9baf326e807
+size 13476836524

last-checkpoint/global_step189/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20fbe4f260c41668ed3f55077bbe711d46060d331a9c1cdc4c1e0d3b4d12c9a3
+size 13476836524

last-checkpoint/global_step189/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:340780e2ee121d44a1a5d36154724dd16398dd1a6afaa0a86d377cd45c46fa91
+size 13476836524

last-checkpoint/global_step189/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:92164a44c14199a14d031be4ec806a60b35861556540e3619dd9ac574c86fbf5
+size 13476836524

last-checkpoint/global_step189/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fad6f8e38bd53cc02696fdd324acefc6450b408b94aa0efc4ff10b7c6725a66e
+size 150693

last-checkpoint/global_step189/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f2769a91ee3073096451825a14033196304aa61466a0742d138c51e7daf74827
+size 150693

last-checkpoint/global_step189/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a1731958171867a7d85e560aa4daebfb823dbd70dd783e2e66bc88bb7340988
+size 150693

last-checkpoint/global_step189/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:65f5c3867311d6afda2cf705c7b21af911b58685f925c8915171e17d8ab9b14d
+size 150693

last-checkpoint/latest

@@ -1 +1 @@
-global_step63
+global_step189

last-checkpoint/model-00001-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fea905c9d1bda6dfe806b7394e5ee3cd1a62a477f8d3a1973c43b3267b722112
+oid sha256:e85b9e60ca25ec6963857f40009ae58911152e3dcccb8a50da8165fded4126ea
 size 4938985352

last-checkpoint/model-00002-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:77a01d7dd5eb5ffe4088b2fcf5f8240025354e4cbd9b8c5f6d69080b39f2e7b7
+oid sha256:ad5dadd7b7b55828763cf8fa416f5592b95088bf1acba10ed355a3cec2a432e7
 size 4947390880

last-checkpoint/model-00003-of-00003.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b5ba87cf8ebe228a95570be373bf45462081878b142dad79ecba9b09eb99b8c9
+oid sha256:356b7afe32f907a9f6e32cd2609fdec01e92fe3d9c34bc8fe6a0542f583fc055
 size 3590488816

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1ce770862b76ceecb3453a5d0b9ef65311eab0f6a60f52997726a89d718de08
+oid sha256:ab7cf0d34d60dfda516cf9661904550e2e294e723edd07c25c738f05e8ba92d1
 size 15024

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2bcae6573442578b752be3e988d4ccb38056b45c31b5c02eb579ec6cebcfa62e
+oid sha256:f7b183324e8227a51a9556d86b2ad893a8c4c52205ed4a737356c6611dac7353
 size 15024

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:398ff45072fee2975e88b9e078f915103d75d94bc08753303cf855b915973623
+oid sha256:ac69e994090f4818cb1fa6f6cefa363178552c3c731c6507ff195bcb07fd5bef
 size 15024

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7f4b210bc832d4c648e3bee8c21dca26e5a8b365d6ec90c638062005a052e57b
+oid sha256:f68e61b63402f8afb1f69c960f7944965655dac11e3ccf29919c282f23931f86
 size 15024

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:845f8eea760ff90e22dadf5fb7f80c9079f8c2298a8770c5d4c3b3309b38a6ef
+oid sha256:c988df4a306242fbc45c908424c576078313f1b5bb472bee13c3b6371b94095e
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 63,
+  "global_step": 189,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -128,6 +128,218 @@
       "eval_samples_per_second": 13.582,
       "eval_steps_per_second": 0.475,
       "step": 63
+    },
+    {
+      "epoch": 1.1111111111111112,
+      "grad_norm": 13.811509421726951,
+      "learning_rate": 3.5e-07,
+      "logits/chosen": -1.53125,
+      "logits/rejected": -1.546875,
+      "logps/chosen": -142.0,
+      "logps/rejected": -147.0,
+      "loss": 0.5957,
+      "rewards/accuracies": 0.675000011920929,
+      "rewards/chosen": -0.1337890625,
+      "rewards/margins": 0.2177734375,
+      "rewards/rejected": -0.3515625,
+      "step": 70
+    },
+    {
+      "epoch": 1.2698412698412698,
+      "grad_norm": 11.993229269242237,
+      "learning_rate": 3.205882352941177e-07,
+      "logits/chosen": -1.5546875,
+      "logits/rejected": -1.5390625,
+      "logps/chosen": -150.0,
+      "logps/rejected": -135.0,
+      "loss": 0.5614,
+      "rewards/accuracies": 0.800000011920929,
+      "rewards/chosen": -0.1328125,
+      "rewards/margins": 0.28515625,
+      "rewards/rejected": -0.41796875,
+      "step": 80
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 11.581245327928885,
+      "learning_rate": 2.911764705882353e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.5078125,
+      "logps/chosen": -144.0,
+      "logps/rejected": -133.0,
+      "loss": 0.5406,
+      "rewards/accuracies": 0.8125,
+      "rewards/chosen": -0.2333984375,
+      "rewards/margins": 0.341796875,
+      "rewards/rejected": -0.57421875,
+      "step": 90
+    },
+    {
+      "epoch": 1.5873015873015874,
+      "grad_norm": 12.825913209383442,
+      "learning_rate": 2.6176470588235295e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.53125,
+      "logps/chosen": -152.0,
+      "logps/rejected": -138.0,
+      "loss": 0.5382,
+      "rewards/accuracies": 0.762499988079071,
+      "rewards/chosen": -0.26953125,
+      "rewards/margins": 0.390625,
+      "rewards/rejected": -0.66015625,
+      "step": 100
+    },
+    {
+      "epoch": 1.746031746031746,
+      "grad_norm": 12.686457389394935,
+      "learning_rate": 2.323529411764706e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.5625,
+      "logps/chosen": -160.0,
+      "logps/rejected": -154.0,
+      "loss": 0.5295,
+      "rewards/accuracies": 0.9125000238418579,
+      "rewards/chosen": -0.23046875,
+      "rewards/margins": 0.46484375,
+      "rewards/rejected": -0.6953125,
+      "step": 110
+    },
+    {
+      "epoch": 1.9047619047619047,
+      "grad_norm": 11.815278327307368,
+      "learning_rate": 2.0294117647058823e-07,
+      "logits/chosen": -1.5546875,
+      "logits/rejected": -1.5,
+      "logps/chosen": -138.0,
+      "logps/rejected": -137.0,
+      "loss": 0.5284,
+      "rewards/accuracies": 0.800000011920929,
+      "rewards/chosen": -0.33984375,
+      "rewards/margins": 0.365234375,
+      "rewards/rejected": -0.70703125,
+      "step": 120
+    },
+    {
+      "epoch": 2.0,
+      "eval_logits/chosen": -1.484375,
+      "eval_logits/rejected": -1.4765625,
+      "eval_logps/chosen": -143.0,
+      "eval_logps/rejected": -130.0,
+      "eval_loss": 0.6310937404632568,
+      "eval_rewards/accuracies": 0.625,
+      "eval_rewards/chosen": -0.388671875,
+      "eval_rewards/margins": 0.185546875,
+      "eval_rewards/rejected": -0.57421875,
+      "eval_runtime": 17.8485,
+      "eval_samples_per_second": 11.205,
+      "eval_steps_per_second": 0.392,
+      "step": 126
+    },
+    {
+      "epoch": 2.0634920634920633,
+      "grad_norm": 10.190155022883303,
+      "learning_rate": 1.7352941176470587e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.515625,
+      "logps/chosen": -154.0,
+      "logps/rejected": -146.0,
+      "loss": 0.4902,
+      "rewards/accuracies": 0.875,
+      "rewards/chosen": -0.298828125,
+      "rewards/margins": 0.51953125,
+      "rewards/rejected": -0.8203125,
+      "step": 130
+    },
+    {
+      "epoch": 2.2222222222222223,
+      "grad_norm": 11.482317016645608,
+      "learning_rate": 1.441176470588235e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.5390625,
+      "logps/chosen": -146.0,
+      "logps/rejected": -147.0,
+      "loss": 0.4618,
+      "rewards/accuracies": 0.862500011920929,
+      "rewards/chosen": -0.298828125,
+      "rewards/margins": 0.57421875,
+      "rewards/rejected": -0.875,
+      "step": 140
+    },
+    {
+      "epoch": 2.380952380952381,
+      "grad_norm": 10.664901408549657,
+      "learning_rate": 1.1470588235294116e-07,
+      "logits/chosen": -1.484375,
+      "logits/rejected": -1.484375,
+      "logps/chosen": -139.0,
+      "logps/rejected": -135.0,
+      "loss": 0.4529,
+      "rewards/accuracies": 0.8999999761581421,
+      "rewards/chosen": -0.392578125,
+      "rewards/margins": 0.56640625,
+      "rewards/rejected": -0.95703125,
+      "step": 150
+    },
+    {
+      "epoch": 2.5396825396825395,
+      "grad_norm": 11.135694541597557,
+      "learning_rate": 8.529411764705883e-08,
+      "logits/chosen": -1.546875,
+      "logits/rejected": -1.46875,
+      "logps/chosen": -138.0,
+      "logps/rejected": -125.5,
+      "loss": 0.4585,
+      "rewards/accuracies": 0.8999999761581421,
+      "rewards/chosen": -0.34375,
+      "rewards/margins": 0.62109375,
+      "rewards/rejected": -0.96484375,
+      "step": 160
+    },
+    {
+      "epoch": 2.6984126984126986,
+      "grad_norm": 9.969592838659052,
+      "learning_rate": 5.5882352941176474e-08,
+      "logits/chosen": -1.515625,
+      "logits/rejected": -1.515625,
+      "logps/chosen": -144.0,
+      "logps/rejected": -147.0,
+      "loss": 0.4698,
+      "rewards/accuracies": 0.8999999761581421,
+      "rewards/chosen": -0.359375,
+      "rewards/margins": 0.60546875,
+      "rewards/rejected": -0.96484375,
+      "step": 170
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 10.412087107819227,
+      "learning_rate": 2.6470588235294116e-08,
+      "logits/chosen": -1.5234375,
+      "logits/rejected": -1.4921875,
+      "logps/chosen": -147.0,
+      "logps/rejected": -137.0,
+      "loss": 0.4529,
+      "rewards/accuracies": 0.9125000238418579,
+      "rewards/chosen": -0.267578125,
+      "rewards/margins": 0.63671875,
+      "rewards/rejected": -0.90234375,
+      "step": 180
+    },
+    {
+      "epoch": 3.0,
+      "eval_logits/chosen": -1.484375,
+      "eval_logits/rejected": -1.46875,
+      "eval_logps/chosen": -143.0,
+      "eval_logps/rejected": -131.0,
+      "eval_loss": 0.6217187643051147,
+      "eval_rewards/accuracies": 0.6785714030265808,
+      "eval_rewards/chosen": -0.462890625,
+      "eval_rewards/margins": 0.220703125,
+      "eval_rewards/rejected": -0.68359375,
+      "eval_runtime": 17.9909,
+      "eval_samples_per_second": 11.117,
+      "eval_steps_per_second": 0.389,
+      "step": 189
     }
   ],
   "logging_steps": 10,
@@ -142,7 +354,7 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": false
+        "should_training_stop": true
       },
       "attributes": {}
     }