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 sha256:59db0ca144f001e1853d14af69a842f5988dc9db8beba7771e981676b8a02139
+size 14496051368

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 sha256:2ee64e731bda9e312733f6afedcd335c7598d998c4ebe6513b0a035c66ed4475
+size 14496051368

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:e5d707226c127f7e3274e226fe2c9bac2da6e90020431d8d82801840d775b87f
+size 14496051368

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 sha256:917861443095935892c255124381352ee996b2a36df1fd278bc8be52b8f162a6
+size 14496051368

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:e64a9c51ebd628b9bf2608cff3d7237a80d5c468cd8d8b48f515124c642a108c
+size 150629

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 sha256:7dd73a544ae58c58529ab5f1c8394bd4ca27de8f03d2c08be023cf8d20238c02
+size 150629

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 sha256:28d28ca8245f2279317f34b723b619870c155e960ec021b9c34c03a1d87d3b4f
+size 150629

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:a1d3e667ee41fda048808b51f84ce0bf14e8f331e4486526d954df9649eeeb8a
+size 150629

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 sha256:462562d909106915dc079f0db44d186f7a516e2648a0204c1db2a6d8b795d3ba
+oid sha256:dd8ab582864f0ff4009bf3594b02ef9b7396e140aa759895b53e77cfc6efe705
 size 4949453792

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:77df233a19e5ee1cf1a9266f75d8a95a933b779063d2045954765829e396d658
+oid sha256:84a95102e3020b65707b3bfd72a5eeca4ad080ec73602c676d1e82d054bc9f74
 size 4999819336

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ef923603dbb36056d914814852707ff18d38f657db76a4b7c95b590dba08a212
+oid sha256:ead6920e38a5285362ca4b0227e6f26923993477c941a50e4d735e7035eedc67
 size 4546807800

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:c019f2bd3222f739db75f458caa9c28765649010b6fa70739851be6f2b00c68b
+oid sha256:c3fd160e69e5766b513d913785299b2ff6d181cae9fd47f741adc100cb17d22e
 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": 14.465,
       "eval_steps_per_second": 0.506,
       "step": 313
+    },
+    {
+      "epoch": 1.0223642172523961,
+      "grad_norm": 0.014031360468919892,
+      "learning_rate": 3.662721893491124e-07,
+      "logits/chosen": -3.140625,
+      "logits/rejected": -3.15625,
+      "logps/chosen": -128.0,
+      "logps/rejected": -330.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.61328125,
+      "rewards/margins": 21.25,
+      "rewards/rejected": -22.0,
+      "step": 320
+    },
+    {
+      "epoch": 1.0543130990415335,
+      "grad_norm": 0.01839801735368299,
+      "learning_rate": 3.603550295857988e-07,
+      "logits/chosen": -3.109375,
+      "logits/rejected": -3.0625,
+      "logps/chosen": -135.0,
+      "logps/rejected": -312.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.5703125,
+      "rewards/margins": 19.75,
+      "rewards/rejected": -21.375,
+      "step": 330
+    },
+    {
+      "epoch": 1.0862619808306708,
+      "grad_norm": 0.016992373953575654,
+      "learning_rate": 3.544378698224852e-07,
+      "logits/chosen": -3.015625,
+      "logits/rejected": -3.0625,
+      "logps/chosen": -105.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.859375,
+      "rewards/margins": 22.625,
+      "rewards/rejected": -23.5,
+      "step": 340
+    },
+    {
+      "epoch": 1.1182108626198084,
+      "grad_norm": 0.00648353239139669,
+      "learning_rate": 3.485207100591716e-07,
+      "logits/chosen": -3.09375,
+      "logits/rejected": -3.0625,
+      "logps/chosen": -134.0,
+      "logps/rejected": -310.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.10009765625,
+      "rewards/margins": 20.5,
+      "rewards/rejected": -20.625,
+      "step": 350
+    },
+    {
+      "epoch": 1.1501597444089458,
+      "grad_norm": 0.018529336855778705,
+      "learning_rate": 3.42603550295858e-07,
+      "logits/chosen": -3.109375,
+      "logits/rejected": -3.0625,
+      "logps/chosen": -121.0,
+      "logps/rejected": -316.0,
+      "loss": 0.0009,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1298828125,
+      "rewards/margins": 21.25,
+      "rewards/rejected": -21.125,
+      "step": 360
+    },
+    {
+      "epoch": 1.182108626198083,
+      "grad_norm": 0.0025056934286525165,
+      "learning_rate": 3.366863905325444e-07,
+      "logits/chosen": -3.15625,
+      "logits/rejected": -3.09375,
+      "logps/chosen": -116.5,
+      "logps/rejected": -314.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.392578125,
+      "rewards/margins": 21.0,
+      "rewards/rejected": -20.5,
+      "step": 370
+    },
+    {
+      "epoch": 1.2140575079872205,
+      "grad_norm": 0.0014546590073500464,
+      "learning_rate": 3.3076923076923075e-07,
+      "logits/chosen": -3.125,
+      "logits/rejected": -3.09375,
+      "logps/chosen": -114.5,
+      "logps/rejected": -318.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.2421875,
+      "rewards/margins": 21.125,
+      "rewards/rejected": -20.875,
+      "step": 380
+    },
+    {
+      "epoch": 1.2460063897763578,
+      "grad_norm": 0.02764778364579726,
+      "learning_rate": 3.2485207100591716e-07,
+      "logits/chosen": -3.078125,
+      "logits/rejected": -3.03125,
+      "logps/chosen": -108.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1494140625,
+      "rewards/margins": 22.875,
+      "rewards/rejected": -23.0,
+      "step": 390
+    },
+    {
+      "epoch": 1.2779552715654952,
+      "grad_norm": 0.002824886589609134,
+      "learning_rate": 3.189349112426035e-07,
+      "logits/chosen": -3.171875,
+      "logits/rejected": -3.15625,
+      "logps/chosen": -147.0,
+      "logps/rejected": -322.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.259765625,
+      "rewards/margins": 21.0,
+      "rewards/rejected": -21.25,
+      "step": 400
+    },
+    {
+      "epoch": 1.3099041533546325,
+      "grad_norm": 0.003952105532040841,
+      "learning_rate": 3.130177514792899e-07,
+      "logits/chosen": -3.125,
+      "logits/rejected": -3.0625,
+      "logps/chosen": -123.5,
+      "logps/rejected": -344.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.201171875,
+      "rewards/margins": 23.875,
+      "rewards/rejected": -23.625,
+      "step": 410
+    },
+    {
+      "epoch": 1.34185303514377,
+      "grad_norm": 0.0010212333678410641,
+      "learning_rate": 3.071005917159763e-07,
+      "logits/chosen": -3.09375,
+      "logits/rejected": -3.109375,
+      "logps/chosen": -128.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1494140625,
+      "rewards/margins": 23.5,
+      "rewards/rejected": -23.75,
+      "step": 420
+    },
+    {
+      "epoch": 1.3738019169329074,
+      "grad_norm": 0.026417806261673966,
+      "learning_rate": 3.011834319526627e-07,
+      "logits/chosen": -3.140625,
+      "logits/rejected": -3.109375,
+      "logps/chosen": -128.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.234375,
+      "rewards/margins": 23.75,
+      "rewards/rejected": -24.0,
+      "step": 430
+    },
+    {
+      "epoch": 1.4057507987220448,
+      "grad_norm": 0.33120053981322833,
+      "learning_rate": 2.952662721893491e-07,
+      "logits/chosen": -3.046875,
+      "logits/rejected": -3.09375,
+      "logps/chosen": -128.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.22265625,
+      "rewards/margins": 24.875,
+      "rewards/rejected": -25.125,
+      "step": 440
+    },
+    {
+      "epoch": 1.4376996805111821,
+      "grad_norm": 5.0669153760203136e-05,
+      "learning_rate": 2.893491124260355e-07,
+      "logits/chosen": -3.03125,
+      "logits/rejected": -3.109375,
+      "logps/chosen": -113.5,
+      "logps/rejected": -348.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0093994140625,
+      "rewards/margins": 23.875,
+      "rewards/rejected": -23.875,
+      "step": 450
+    },
+    {
+      "epoch": 1.4696485623003195,
+      "grad_norm": 0.002235783476143995,
+      "learning_rate": 2.834319526627219e-07,
+      "logits/chosen": -3.046875,
+      "logits/rejected": -3.015625,
+      "logps/chosen": -122.5,
+      "logps/rejected": -346.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.42578125,
+      "rewards/margins": 23.625,
+      "rewards/rejected": -24.0,
+      "step": 460
+    },
+    {
+      "epoch": 1.5015974440894568,
+      "grad_norm": 0.0021115625525813,
+      "learning_rate": 2.7751479289940824e-07,
+      "logits/chosen": -3.0625,
+      "logits/rejected": -3.046875,
+      "logps/chosen": -127.5,
+      "logps/rejected": -352.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.2490234375,
+      "rewards/margins": 24.5,
+      "rewards/rejected": -24.75,
+      "step": 470
+    },
+    {
+      "epoch": 1.5335463258785942,
+      "grad_norm": 0.0005420939322641936,
+      "learning_rate": 2.715976331360947e-07,
+      "logits/chosen": -3.09375,
+      "logits/rejected": -3.015625,
+      "logps/chosen": -148.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8828125,
+      "rewards/margins": 24.125,
+      "rewards/rejected": -25.0,
+      "step": 480
+    },
+    {
+      "epoch": 1.5654952076677318,
+      "grad_norm": 0.007358687931687971,
+      "learning_rate": 2.6568047337278106e-07,
+      "logits/chosen": -3.046875,
+      "logits/rejected": -3.046875,
+      "logps/chosen": -125.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.625,
+      "rewards/margins": 24.75,
+      "rewards/rejected": -25.375,
+      "step": 490
+    },
+    {
+      "epoch": 1.5974440894568689,
+      "grad_norm": 0.0654742270230356,
+      "learning_rate": 2.5976331360946746e-07,
+      "logits/chosen": -3.0625,
+      "logits/rejected": -3.046875,
+      "logps/chosen": -124.0,
+      "logps/rejected": -334.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -1.046875,
+      "rewards/margins": 21.875,
+      "rewards/rejected": -22.875,
+      "step": 500
+    },
+    {
+      "epoch": 1.6293929712460065,
+      "grad_norm": 0.0008641901527854177,
+      "learning_rate": 2.538461538461538e-07,
+      "logits/chosen": -3.0625,
+      "logits/rejected": -3.078125,
+      "logps/chosen": -131.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.69140625,
+      "rewards/margins": 22.875,
+      "rewards/rejected": -23.625,
+      "step": 510
+    },
+    {
+      "epoch": 1.6613418530351438,
+      "grad_norm": 0.015004697710408908,
+      "learning_rate": 2.479289940828402e-07,
+      "logits/chosen": -3.015625,
+      "logits/rejected": -3.0625,
+      "logps/chosen": -100.5,
+      "logps/rejected": -374.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.0927734375,
+      "rewards/margins": 26.5,
+      "rewards/rejected": -26.625,
+      "step": 520
+    },
+    {
+      "epoch": 1.6932907348242812,
+      "grad_norm": 0.01615930039253711,
+      "learning_rate": 2.420118343195266e-07,
+      "logits/chosen": -3.0625,
+      "logits/rejected": -3.0,
+      "logps/chosen": -130.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.010009765625,
+      "rewards/margins": 23.25,
+      "rewards/rejected": -23.25,
+      "step": 530
+    },
+    {
+      "epoch": 1.7252396166134185,
+      "grad_norm": 0.005625013393470251,
+      "learning_rate": 2.36094674556213e-07,
+      "logits/chosen": -3.015625,
+      "logits/rejected": -3.015625,
+      "logps/chosen": -123.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.482421875,
+      "rewards/margins": 23.375,
+      "rewards/rejected": -22.875,
+      "step": 540
+    },
+    {
+      "epoch": 1.7571884984025559,
+      "grad_norm": 0.006628504166462446,
+      "learning_rate": 2.301775147928994e-07,
+      "logits/chosen": -3.109375,
+      "logits/rejected": -3.0625,
+      "logps/chosen": -128.0,
+      "logps/rejected": -324.0,
+      "loss": 0.0008,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.1650390625,
+      "rewards/margins": 21.75,
+      "rewards/rejected": -21.5,
+      "step": 550
+    },
+    {
+      "epoch": 1.7891373801916934,
+      "grad_norm": 0.0051543594641130655,
+      "learning_rate": 2.242603550295858e-07,
+      "logits/chosen": -3.0625,
+      "logits/rejected": -3.0,
+      "logps/chosen": -102.0,
+      "logps/rejected": -336.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.283203125,
+      "rewards/margins": 23.25,
+      "rewards/rejected": -23.0,
+      "step": 560
+    },
+    {
+      "epoch": 1.8210862619808306,
+      "grad_norm": 0.0019828665244342454,
+      "learning_rate": 2.1834319526627217e-07,
+      "logits/chosen": -2.890625,
+      "logits/rejected": -2.9375,
+      "logps/chosen": -112.0,
+      "logps/rejected": -328.0,
+      "loss": 0.001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.051513671875,
+      "rewards/margins": 22.375,
+      "rewards/rejected": -22.375,
+      "step": 570
+    },
+    {
+      "epoch": 1.8530351437699681,
+      "grad_norm": 0.0010506656519779034,
+      "learning_rate": 2.1242603550295858e-07,
+      "logits/chosen": -3.03125,
+      "logits/rejected": -2.953125,
+      "logps/chosen": -133.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.11767578125,
+      "rewards/margins": 23.125,
+      "rewards/rejected": -23.25,
+      "step": 580
+    },
+    {
+      "epoch": 1.8849840255591053,
+      "grad_norm": 0.026254564073093664,
+      "learning_rate": 2.0650887573964496e-07,
+      "logits/chosen": -2.9375,
+      "logits/rejected": -2.921875,
+      "logps/chosen": -126.0,
+      "logps/rejected": -332.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1484375,
+      "rewards/margins": 22.0,
+      "rewards/rejected": -22.125,
+      "step": 590
+    },
+    {
+      "epoch": 1.9169329073482428,
+      "grad_norm": 0.0011075869246766719,
+      "learning_rate": 2.0059171597633133e-07,
+      "logits/chosen": -2.921875,
+      "logits/rejected": -2.875,
+      "logps/chosen": -97.0,
+      "logps/rejected": -326.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 0.12109375,
+      "rewards/margins": 22.375,
+      "rewards/rejected": -22.25,
+      "step": 600
+    },
+    {
+      "epoch": 1.9488817891373802,
+      "grad_norm": 3.965796587839269,
+      "learning_rate": 1.9467455621301774e-07,
+      "logits/chosen": -2.9375,
+      "logits/rejected": -2.890625,
+      "logps/chosen": -140.0,
+      "logps/rejected": -338.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.404296875,
+      "rewards/margins": 22.75,
+      "rewards/rejected": -23.125,
+      "step": 610
+    },
+    {
+      "epoch": 1.9808306709265175,
+      "grad_norm": 0.14930442211812853,
+      "learning_rate": 1.8875739644970412e-07,
+      "logits/chosen": -2.921875,
+      "logits/rejected": -2.890625,
+      "logps/chosen": -116.5,
+      "logps/rejected": -370.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.9765625,
+      "rewards/margins": 25.25,
+      "rewards/rejected": -26.25,
+      "step": 620
+    },
+    {
+      "epoch": 2.0,
+      "eval_logits/chosen": -3.03125,
+      "eval_logits/rejected": -2.953125,
+      "eval_logps/chosen": -175.0,
+      "eval_logps/rejected": -336.0,
+      "eval_loss": 0.0031948494724929333,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": -0.7578125,
+      "eval_rewards/margins": 21.375,
+      "eval_rewards/rejected": -22.125,
+      "eval_runtime": 13.9279,
+      "eval_samples_per_second": 14.36,
+      "eval_steps_per_second": 0.503,
+      "step": 626
     }
   ],
   "logging_steps": 10,