Training in progress, epoch 2, checkpoint

last-checkpoint/global_step626/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:bbf5442bcada9db1fde4475cbc0bf5a603ebc18aa1b46251961608fe3e0dbed8
+size 13476836524

last-checkpoint/global_step626/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:109b55dc8e802c0e56485448608105d657d60cc9e059519c2ea800582cb0dc1a
+size 13476836524

last-checkpoint/global_step626/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:27aa416c955e1423f16c6269698a43b5c9f8ff10116732bf7cb40e95bebe55a5
+size 13476836524

last-checkpoint/global_step626/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:f8ea065358139c23c16e22e695efb0ff84ecd09118284a02707dbb86ed1fe71a
+size 13476836524

last-checkpoint/global_step626/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f04249e1b8fe1c2cdc217cf32f7eea80a19cb28668041a7568bda12381df332
+size 150693

last-checkpoint/global_step626/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2a8e7204e992a56a415d70b4b33a5427de747e33a8410a8d2e57b2d3760f41a3
+size 150693

last-checkpoint/global_step626/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b82bc0de7fa407ebf2e63d8203a63e23579c6341814abb2367c0d8f0fb95b0c2
+size 150693

last-checkpoint/global_step626/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f568cb4b0ff22d0f18efb7a633cdedc227f3a80be6b0c1fcc6c09a4985216c2c
+size 150693

last-checkpoint/latest

@@ -1 +1 @@
-global_step313
+global_step626

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dba48ff426d1d3f62349fb0828ec2ce1a0bcdd0d3e73f0bdbcd91d85047b36ac
+oid sha256:37d98d0d79a2a8056e888825dbb89319bd37cc90e78333311f476731c0de96dd
 size 4938985352

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:09b39e4252b18ddce9733e44c268d3b7ea8083d7119c409a7cdd98507616733d
+oid sha256:98a0d6dc670bcc211a513dffddee9dab1866d26bfe1ddbab630b87c3d20efcb6
 size 4947390880

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52f59febeb0d35b9b4bb39aa39799a5b0fe5d26be6e53e9ba1065a3a6fe1b1e5
+oid sha256:ac732abca0dd08baaa3ac8e6c052d653eb7d4789e9b99246e8456185f80cbf57
 size 3590488816

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f1ce770862b76ceecb3453a5d0b9ef65311eab0f6a60f52997726a89d718de08
+oid sha256:2307c03867cef25b5028feb9a23f80e784b9af9a615de13ddca560a6a90fb593
 size 15024

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2bcae6573442578b752be3e988d4ccb38056b45c31b5c02eb579ec6cebcfa62e
+oid sha256:50218cfaecdd818354e567b7167c13899e3b42297e7d8f58bd7e732cfa547800
 size 15024

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:398ff45072fee2975e88b9e078f915103d75d94bc08753303cf855b915973623
+oid sha256:9814a66b49861f5495b06dae3be12ddf7185b88e2cae1fb808ca9efd99d5807f
 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:a7be93040a65e0a29975f6c70b94418e1fdf88423a50c58aa572141d3c92fbfc
 size 15024

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6b9aad3a748e15719e20120f0a760f22686338e176ee3f1c011b6dab2f3ab19b
+oid sha256:f7725b44c69725321786f1e58dccd7ea4d3ae5794ea47bd49c0d4a139aec266d
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 2.0,
   "eval_steps": 500,
-  "global_step": 313,
+  "global_step": 626,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -503,6 +503,487 @@
       "eval_samples_per_second": 10.003,
       "eval_steps_per_second": 0.35,
       "step": 313
+    },
+    {
+      "epoch": 1.0223642172523961,
+      "grad_norm": 0.022677077708026305,
+      "learning_rate": 3.662721893491124e-07,
+      "logits/chosen": -0.73046875,
+      "logits/rejected": -0.44140625,
+      "logps/chosen": -278.0,
+      "logps/rejected": -240.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.431640625,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.1875,
+      "step": 320
+    },
+    {
+      "epoch": 1.0543130990415335,
+      "grad_norm": 0.011840880769205434,
+      "learning_rate": 3.603550295857988e-07,
+      "logits/chosen": -0.72265625,
+      "logits/rejected": -0.5546875,
+      "logps/chosen": -322.0,
+      "logps/rejected": -240.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.380859375,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.375,
+      "step": 330
+    },
+    {
+      "epoch": 1.0862619808306708,
+      "grad_norm": 0.00391406917206777,
+      "learning_rate": 3.544378698224852e-07,
+      "logits/chosen": -0.671875,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -314.0,
+      "logps/rejected": -245.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.359375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.5625,
+      "step": 340
+    },
+    {
+      "epoch": 1.1182108626198084,
+      "grad_norm": 0.0036384946886406336,
+      "learning_rate": 3.485207100591716e-07,
+      "logits/chosen": -0.7265625,
+      "logits/rejected": -0.5546875,
+      "logps/chosen": -314.0,
+      "logps/rejected": -235.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.263671875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.25,
+      "step": 350
+    },
+    {
+      "epoch": 1.1501597444089458,
+      "grad_norm": 0.35575156394070306,
+      "learning_rate": 3.42603550295858e-07,
+      "logits/chosen": -0.73828125,
+      "logits/rejected": -0.546875,
+      "logps/chosen": -274.0,
+      "logps/rejected": -246.0,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.69140625,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.6875,
+      "step": 360
+    },
+    {
+      "epoch": 1.182108626198083,
+      "grad_norm": 0.4647357188457555,
+      "learning_rate": 3.366863905325444e-07,
+      "logits/chosen": -0.58203125,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -251.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.384765625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -12.625,
+      "step": 370
+    },
+    {
+      "epoch": 1.2140575079872205,
+      "grad_norm": 0.006194595470527809,
+      "learning_rate": 3.3076923076923075e-07,
+      "logits/chosen": -0.6875,
+      "logits/rejected": -0.373046875,
+      "logps/chosen": -284.0,
+      "logps/rejected": -240.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5234375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.6875,
+      "step": 380
+    },
+    {
+      "epoch": 1.2460063897763578,
+      "grad_norm": 0.08957807843959086,
+      "learning_rate": 3.2485207100591716e-07,
+      "logits/chosen": -0.7265625,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -342.0,
+      "logps/rejected": -245.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5859375,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -12.75,
+      "step": 390
+    },
+    {
+      "epoch": 1.2779552715654952,
+      "grad_norm": 0.0040259036143305595,
+      "learning_rate": 3.189349112426035e-07,
+      "logits/chosen": -0.65625,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -324.0,
+      "logps/rejected": -248.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.451171875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.0,
+      "step": 400
+    },
+    {
+      "epoch": 1.3099041533546325,
+      "grad_norm": 0.007242003331847797,
+      "learning_rate": 3.130177514792899e-07,
+      "logits/chosen": -0.6484375,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -318.0,
+      "logps/rejected": -248.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.61328125,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.5,
+      "step": 410
+    },
+    {
+      "epoch": 1.34185303514377,
+      "grad_norm": 0.0064360979279180575,
+      "learning_rate": 3.071005917159763e-07,
+      "logits/chosen": -0.58203125,
+      "logits/rejected": -0.5625,
+      "logps/chosen": -296.0,
+      "logps/rejected": -249.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.03125,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -13.125,
+      "step": 420
+    },
+    {
+      "epoch": 1.3738019169329074,
+      "grad_norm": 0.00975201883790689,
+      "learning_rate": 3.011834319526627e-07,
+      "logits/chosen": -0.64453125,
+      "logits/rejected": -0.51953125,
+      "logps/chosen": -314.0,
+      "logps/rejected": -252.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.64453125,
+      "rewards/margins": 12.6875,
+      "rewards/rejected": -13.375,
+      "step": 430
+    },
+    {
+      "epoch": 1.4057507987220448,
+      "grad_norm": 0.005359836815678541,
+      "learning_rate": 2.952662721893491e-07,
+      "logits/chosen": -0.65234375,
+      "logits/rejected": -0.4609375,
+      "logps/chosen": -336.0,
+      "logps/rejected": -249.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.69140625,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.5,
+      "step": 440
+    },
+    {
+      "epoch": 1.4376996805111821,
+      "grad_norm": 0.07525889423766378,
+      "learning_rate": 2.893491124260355e-07,
+      "logits/chosen": -0.7109375,
+      "logits/rejected": -0.5546875,
+      "logps/chosen": -340.0,
+      "logps/rejected": -254.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75,
+      "rewards/margins": 12.875,
+      "rewards/rejected": -13.625,
+      "step": 450
+    },
+    {
+      "epoch": 1.4696485623003195,
+      "grad_norm": 0.00193288934052946,
+      "learning_rate": 2.834319526627219e-07,
+      "logits/chosen": -0.703125,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -300.0,
+      "logps/rejected": -255.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8125,
+      "rewards/margins": 12.8125,
+      "rewards/rejected": -13.6875,
+      "step": 460
+    },
+    {
+      "epoch": 1.5015974440894568,
+      "grad_norm": 0.004933186728273418,
+      "learning_rate": 2.7751479289940824e-07,
+      "logits/chosen": -0.64453125,
+      "logits/rejected": -0.453125,
+      "logps/chosen": -294.0,
+      "logps/rejected": -252.0,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.796875,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.5625,
+      "step": 470
+    },
+    {
+      "epoch": 1.5335463258785942,
+      "grad_norm": 0.09834967775177374,
+      "learning_rate": 2.715976331360947e-07,
+      "logits/chosen": -0.62890625,
+      "logits/rejected": -0.478515625,
+      "logps/chosen": -360.0,
+      "logps/rejected": -247.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7109375,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.25,
+      "step": 480
+    },
+    {
+      "epoch": 1.5654952076677318,
+      "grad_norm": 0.1180240734795196,
+      "learning_rate": 2.6568047337278106e-07,
+      "logits/chosen": -0.69921875,
+      "logits/rejected": -0.5078125,
+      "logps/chosen": -346.0,
+      "logps/rejected": -258.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.015625,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -14.125,
+      "step": 490
+    },
+    {
+      "epoch": 1.5974440894568689,
+      "grad_norm": 0.10115647498128895,
+      "learning_rate": 2.5976331360946746e-07,
+      "logits/chosen": -0.66015625,
+      "logits/rejected": -0.54296875,
+      "logps/chosen": -372.0,
+      "logps/rejected": -256.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.671875,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -13.625,
+      "step": 500
+    },
+    {
+      "epoch": 1.6293929712460065,
+      "grad_norm": 0.005594304401420574,
+      "learning_rate": 2.538461538461538e-07,
+      "logits/chosen": -0.59375,
+      "logits/rejected": -0.52734375,
+      "logps/chosen": -312.0,
+      "logps/rejected": -258.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.79296875,
+      "rewards/margins": 13.4375,
+      "rewards/rejected": -14.25,
+      "step": 510
+    },
+    {
+      "epoch": 1.6613418530351438,
+      "grad_norm": 0.0027735638022890357,
+      "learning_rate": 2.479289940828402e-07,
+      "logits/chosen": -0.6953125,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -292.0,
+      "logps/rejected": -250.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1875,
+      "rewards/margins": 12.75,
+      "rewards/rejected": -13.9375,
+      "step": 520
+    },
+    {
+      "epoch": 1.6932907348242812,
+      "grad_norm": 0.0005941899982621629,
+      "learning_rate": 2.420118343195266e-07,
+      "logits/chosen": -0.64453125,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -298.0,
+      "logps/rejected": -262.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.9765625,
+      "rewards/margins": 13.3125,
+      "rewards/rejected": -14.3125,
+      "step": 530
+    },
+    {
+      "epoch": 1.7252396166134185,
+      "grad_norm": 0.024572757961485794,
+      "learning_rate": 2.36094674556213e-07,
+      "logits/chosen": -0.72265625,
+      "logits/rejected": -0.51171875,
+      "logps/chosen": -366.0,
+      "logps/rejected": -256.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.046875,
+      "rewards/margins": 12.9375,
+      "rewards/rejected": -14.0,
+      "step": 540
+    },
+    {
+      "epoch": 1.7571884984025559,
+      "grad_norm": 0.0008800867821762263,
+      "learning_rate": 2.301775147928994e-07,
+      "logits/chosen": -0.6875,
+      "logits/rejected": -0.44921875,
+      "logps/chosen": -308.0,
+      "logps/rejected": -256.0,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.89453125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -13.9375,
+      "step": 550
+    },
+    {
+      "epoch": 1.7891373801916934,
+      "grad_norm": 0.001437566874272442,
+      "learning_rate": 2.242603550295858e-07,
+      "logits/chosen": -0.76171875,
+      "logits/rejected": -0.58984375,
+      "logps/chosen": -376.0,
+      "logps/rejected": -260.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.78125,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -13.875,
+      "step": 560
+    },
+    {
+      "epoch": 1.8210862619808306,
+      "grad_norm": 0.0026793717529034755,
+      "learning_rate": 2.1834319526627217e-07,
+      "logits/chosen": -0.6953125,
+      "logits/rejected": -0.5234375,
+      "logps/chosen": -308.0,
+      "logps/rejected": -253.0,
+      "loss": 0.0014,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.98828125,
+      "rewards/margins": 13.1875,
+      "rewards/rejected": -14.1875,
+      "step": 570
+    },
+    {
+      "epoch": 1.8530351437699681,
+      "grad_norm": 0.04424802361114257,
+      "learning_rate": 2.1242603550295858e-07,
+      "logits/chosen": -0.734375,
+      "logits/rejected": -0.56640625,
+      "logps/chosen": -330.0,
+      "logps/rejected": -262.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.1953125,
+      "rewards/margins": 13.0,
+      "rewards/rejected": -14.1875,
+      "step": 580
+    },
+    {
+      "epoch": 1.8849840255591053,
+      "grad_norm": 0.0015499356598387029,
+      "learning_rate": 2.0650887573964496e-07,
+      "logits/chosen": -0.70703125,
+      "logits/rejected": -0.5859375,
+      "logps/chosen": -380.0,
+      "logps/rejected": -262.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.0703125,
+      "rewards/margins": 13.25,
+      "rewards/rejected": -14.3125,
+      "step": 590
+    },
+    {
+      "epoch": 1.9169329073482428,
+      "grad_norm": 0.0017185011208288798,
+      "learning_rate": 2.0059171597633133e-07,
+      "logits/chosen": -0.625,
+      "logits/rejected": -0.498046875,
+      "logps/chosen": -320.0,
+      "logps/rejected": -270.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.015625,
+      "rewards/margins": 13.375,
+      "rewards/rejected": -14.375,
+      "step": 600
+    },
+    {
+      "epoch": 1.9488817891373802,
+      "grad_norm": 0.06027517128383722,
+      "learning_rate": 1.9467455621301774e-07,
+      "logits/chosen": -0.609375,
+      "logits/rejected": -0.443359375,
+      "logps/chosen": -324.0,
+      "logps/rejected": -254.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.96875,
+      "rewards/margins": 13.125,
+      "rewards/rejected": -14.125,
+      "step": 610
+    },
+    {
+      "epoch": 1.9808306709265175,
+      "grad_norm": 0.0034706962600456605,
+      "learning_rate": 1.8875739644970412e-07,
+      "logits/chosen": -0.75390625,
+      "logits/rejected": -0.6015625,
+      "logps/chosen": -338.0,
+      "logps/rejected": -262.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.15625,
+      "rewards/margins": 13.75,
+      "rewards/rejected": -14.875,
+      "step": 620
+    },
+    {
+      "epoch": 2.0,
+      "eval_logits/chosen": -0.71484375,
+      "eval_logits/rejected": -0.4375,
+      "eval_logps/chosen": -304.0,
+      "eval_logps/rejected": -258.0,
+      "eval_loss": 0.00012740046076942235,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": -1.1875,
+      "eval_rewards/margins": 13.25,
+      "eval_rewards/rejected": -14.4375,
+      "eval_runtime": 20.0785,
+      "eval_samples_per_second": 9.961,
+      "eval_steps_per_second": 0.349,
+      "step": 626
     }
   ],
   "logging_steps": 10,