Training in progress, epoch 3, checkpoint

last-checkpoint/global_step939/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:76e580faad10ea668c0da1b8daeb4d1ee215cb511b62ad7fdc876d61354b7604
+size 14496051368

last-checkpoint/global_step939/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:bba26e78d4b2c286ef67ebe56455fafb7e28b6fa8ce9a999ab8bd151512fa0b2
+size 14496051368

last-checkpoint/global_step939/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0207c0724bbca9b56af6d3b95693667d7dd2c8ce8ddf323dd439ea90ff8d1f8a
+size 14496051368

last-checkpoint/global_step939/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4a16fe7e809a68305891c48ceeadfc739f27a3ff0d2de47b3f54bad00b3677bd
+size 14496051368

last-checkpoint/global_step939/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2efe7d77b4410e6649a35dc6efabb73a0dfc5153beda8cefa2495c69bf65fbe
+size 150629

last-checkpoint/global_step939/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3e6cad11b1fd5787adcea9f18e4c482897157146658af7358b8e1330c5de61c5
+size 150629

last-checkpoint/global_step939/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6eb281386924df332f5956c9354cf760b10785f53e4738767a85453249c0b725
+size 150629

last-checkpoint/global_step939/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:32f520b4a0c187f92c6bf6c96396edb2cd0df53e8944e38b0ce6f06db6b59b05
+size 150629

last-checkpoint/latest

@@ -1 +1 @@
-global_step626
+global_step939

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dd8ab582864f0ff4009bf3594b02ef9b7396e140aa759895b53e77cfc6efe705
+oid sha256:f98b6ce9486fcf92a7e44e1f587e0de3a38f28a9aff1a5ecc93ccf817a453db6
 size 4949453792

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:84a95102e3020b65707b3bfd72a5eeca4ad080ec73602c676d1e82d054bc9f74
+oid sha256:23ae631c1839a744dc9965b39b5ca5e6e4b783311f91b4409d3c37dba637f2c5
 size 4999819336

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ead6920e38a5285362ca4b0227e6f26923993477c941a50e4d735e7035eedc67
+oid sha256:94b7b77a1e74f57b84c3f159edc99b1b289a5d6d44427213a5a3b151f65c3443
 size 4546807800

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2307c03867cef25b5028feb9a23f80e784b9af9a615de13ddca560a6a90fb593
+oid sha256:ab7cf0d34d60dfda516cf9661904550e2e294e723edd07c25c738f05e8ba92d1
 size 15024

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:50218cfaecdd818354e567b7167c13899e3b42297e7d8f58bd7e732cfa547800
+oid sha256:f7b183324e8227a51a9556d86b2ad893a8c4c52205ed4a737356c6611dac7353
 size 15024

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9814a66b49861f5495b06dae3be12ddf7185b88e2cae1fb808ca9efd99d5807f
+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:a7be93040a65e0a29975f6c70b94418e1fdf88423a50c58aa572141d3c92fbfc
+oid sha256:f68e61b63402f8afb1f69c960f7944965655dac11e3ccf29919c282f23931f86
 size 15024

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c3fd160e69e5766b513d913785299b2ff6d181cae9fd47f741adc100cb17d22e
+oid sha256:640747ca69b57ef58ae079eec14158d13652c8bf30fb7c733fab30d65edaeb50
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 2.0,
+  "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 626,
+  "global_step": 939,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -984,6 +984,487 @@
       "eval_samples_per_second": 14.36,
       "eval_steps_per_second": 0.503,
       "step": 626
+    },
+    {
+      "epoch": 2.012779552715655,
+      "grad_norm": 0.0032974843044556753,
+      "learning_rate": 1.8284023668639053e-07,
+      "logits/chosen": -2.859375,
+      "logits/rejected": -2.84375,
+      "logps/chosen": -133.0,
+      "logps/rejected": -346.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.486328125,
+      "rewards/margins": 23.5,
+      "rewards/rejected": -24.0,
+      "step": 630
+    },
+    {
+      "epoch": 2.0447284345047922,
+      "grad_norm": 0.1432508053581138,
+      "learning_rate": 1.7692307692307693e-07,
+      "logits/chosen": -2.96875,
+      "logits/rejected": -2.9375,
+      "logps/chosen": -135.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.39453125,
+      "rewards/margins": 22.875,
+      "rewards/rejected": -23.25,
+      "step": 640
+    },
+    {
+      "epoch": 2.07667731629393,
+      "grad_norm": 0.0001731085607822556,
+      "learning_rate": 1.710059171597633e-07,
+      "logits/chosen": -2.765625,
+      "logits/rejected": -2.796875,
+      "logps/chosen": -109.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.19921875,
+      "rewards/margins": 26.625,
+      "rewards/rejected": -26.75,
+      "step": 650
+    },
+    {
+      "epoch": 2.108626198083067,
+      "grad_norm": 0.0019185996024299748,
+      "learning_rate": 1.650887573964497e-07,
+      "logits/chosen": -2.859375,
+      "logits/rejected": -2.8125,
+      "logps/chosen": -122.0,
+      "logps/rejected": -342.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.77734375,
+      "rewards/margins": 23.0,
+      "rewards/rejected": -23.75,
+      "step": 660
+    },
+    {
+      "epoch": 2.1405750798722045,
+      "grad_norm": 0.0008784455055030455,
+      "learning_rate": 1.591715976331361e-07,
+      "logits/chosen": -2.859375,
+      "logits/rejected": -2.859375,
+      "logps/chosen": -125.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.478515625,
+      "rewards/margins": 24.0,
+      "rewards/rejected": -24.5,
+      "step": 670
+    },
+    {
+      "epoch": 2.1725239616613417,
+      "grad_norm": 0.009325179732468034,
+      "learning_rate": 1.5325443786982248e-07,
+      "logits/chosen": -2.828125,
+      "logits/rejected": -2.828125,
+      "logps/chosen": -126.5,
+      "logps/rejected": -358.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5234375,
+      "rewards/margins": 24.5,
+      "rewards/rejected": -25.0,
+      "step": 680
+    },
+    {
+      "epoch": 2.2044728434504792,
+      "grad_norm": 0.0007194651213122703,
+      "learning_rate": 1.4733727810650885e-07,
+      "logits/chosen": -2.9375,
+      "logits/rejected": -2.859375,
+      "logps/chosen": -158.0,
+      "logps/rejected": -362.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.90625,
+      "rewards/margins": 23.875,
+      "rewards/rejected": -24.875,
+      "step": 690
+    },
+    {
+      "epoch": 2.236421725239617,
+      "grad_norm": 0.008732155259144234,
+      "learning_rate": 1.4142011834319526e-07,
+      "logits/chosen": -2.78125,
+      "logits/rejected": -2.765625,
+      "logps/chosen": -126.5,
+      "logps/rejected": -368.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.765625,
+      "rewards/margins": 25.625,
+      "rewards/rejected": -26.375,
+      "step": 700
+    },
+    {
+      "epoch": 2.268370607028754,
+      "grad_norm": 0.0005971006192250428,
+      "learning_rate": 1.3550295857988164e-07,
+      "logits/chosen": -2.8125,
+      "logits/rejected": -2.875,
+      "logps/chosen": -114.5,
+      "logps/rejected": -344.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.46875,
+      "rewards/margins": 23.125,
+      "rewards/rejected": -23.625,
+      "step": 710
+    },
+    {
+      "epoch": 2.3003194888178915,
+      "grad_norm": 0.00013779482104364402,
+      "learning_rate": 1.2958579881656802e-07,
+      "logits/chosen": -2.828125,
+      "logits/rejected": -2.859375,
+      "logps/chosen": -134.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5234375,
+      "rewards/margins": 24.875,
+      "rewards/rejected": -25.375,
+      "step": 720
+    },
+    {
+      "epoch": 2.3322683706070286,
+      "grad_norm": 6.075224878814938e-05,
+      "learning_rate": 1.2366863905325443e-07,
+      "logits/chosen": -2.796875,
+      "logits/rejected": -2.828125,
+      "logps/chosen": -121.0,
+      "logps/rejected": -358.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8046875,
+      "rewards/margins": 24.125,
+      "rewards/rejected": -25.0,
+      "step": 730
+    },
+    {
+      "epoch": 2.364217252396166,
+      "grad_norm": 0.0007165926429755341,
+      "learning_rate": 1.1775147928994082e-07,
+      "logits/chosen": -2.921875,
+      "logits/rejected": -2.8125,
+      "logps/chosen": -137.0,
+      "logps/rejected": -374.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.498046875,
+      "rewards/margins": 26.125,
+      "rewards/rejected": -26.625,
+      "step": 740
+    },
+    {
+      "epoch": 2.3961661341853033,
+      "grad_norm": 0.00026410738853004405,
+      "learning_rate": 1.1183431952662721e-07,
+      "logits/chosen": -2.796875,
+      "logits/rejected": -2.8125,
+      "logps/chosen": -146.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.96875,
+      "rewards/margins": 24.75,
+      "rewards/rejected": -25.75,
+      "step": 750
+    },
+    {
+      "epoch": 2.428115015974441,
+      "grad_norm": 0.02405558435283539,
+      "learning_rate": 1.059171597633136e-07,
+      "logits/chosen": -2.78125,
+      "logits/rejected": -2.8125,
+      "logps/chosen": -143.0,
+      "logps/rejected": -368.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.82421875,
+      "rewards/margins": 24.875,
+      "rewards/rejected": -25.75,
+      "step": 760
+    },
+    {
+      "epoch": 2.460063897763578,
+      "grad_norm": 0.00034653212382964406,
+      "learning_rate": 1e-07,
+      "logits/chosen": -2.84375,
+      "logits/rejected": -2.84375,
+      "logps/chosen": -126.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.69921875,
+      "rewards/margins": 25.625,
+      "rewards/rejected": -26.25,
+      "step": 770
+    },
+    {
+      "epoch": 2.4920127795527156,
+      "grad_norm": 0.0015560065245145238,
+      "learning_rate": 9.408284023668639e-08,
+      "logits/chosen": -2.78125,
+      "logits/rejected": -2.796875,
+      "logps/chosen": -128.0,
+      "logps/rejected": -344.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.51171875,
+      "rewards/margins": 23.375,
+      "rewards/rejected": -23.875,
+      "step": 780
+    },
+    {
+      "epoch": 2.523961661341853,
+      "grad_norm": 0.00033309018788403783,
+      "learning_rate": 8.816568047337278e-08,
+      "logits/chosen": -2.921875,
+      "logits/rejected": -2.859375,
+      "logps/chosen": -134.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.357421875,
+      "rewards/margins": 25.75,
+      "rewards/rejected": -26.0,
+      "step": 790
+    },
+    {
+      "epoch": 2.5559105431309903,
+      "grad_norm": 0.0025148503211950596,
+      "learning_rate": 8.224852071005916e-08,
+      "logits/chosen": -2.78125,
+      "logits/rejected": -2.78125,
+      "logps/chosen": -104.5,
+      "logps/rejected": -382.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.470703125,
+      "rewards/margins": 27.0,
+      "rewards/rejected": -27.5,
+      "step": 800
+    },
+    {
+      "epoch": 2.587859424920128,
+      "grad_norm": 0.00048763191845162235,
+      "learning_rate": 7.633136094674555e-08,
+      "logits/chosen": -2.90625,
+      "logits/rejected": -2.84375,
+      "logps/chosen": -139.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.53125,
+      "rewards/margins": 25.875,
+      "rewards/rejected": -26.375,
+      "step": 810
+    },
+    {
+      "epoch": 2.619808306709265,
+      "grad_norm": 0.0010336427914530199,
+      "learning_rate": 7.041420118343195e-08,
+      "logits/chosen": -2.796875,
+      "logits/rejected": -2.84375,
+      "logps/chosen": -119.0,
+      "logps/rejected": -372.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.435546875,
+      "rewards/margins": 25.75,
+      "rewards/rejected": -26.25,
+      "step": 820
+    },
+    {
+      "epoch": 2.6517571884984026,
+      "grad_norm": 9.273196336389341e-05,
+      "learning_rate": 6.449704142011835e-08,
+      "logits/chosen": -2.765625,
+      "logits/rejected": -2.765625,
+      "logps/chosen": -118.5,
+      "logps/rejected": -366.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.271484375,
+      "rewards/margins": 25.625,
+      "rewards/rejected": -25.875,
+      "step": 830
+    },
+    {
+      "epoch": 2.68370607028754,
+      "grad_norm": 0.0004326121515741712,
+      "learning_rate": 5.857988165680473e-08,
+      "logits/chosen": -2.828125,
+      "logits/rejected": -2.75,
+      "logps/chosen": -139.0,
+      "logps/rejected": -356.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5703125,
+      "rewards/margins": 24.375,
+      "rewards/rejected": -25.0,
+      "step": 840
+    },
+    {
+      "epoch": 2.7156549520766773,
+      "grad_norm": 0.0013592593922840503,
+      "learning_rate": 5.266272189349112e-08,
+      "logits/chosen": -2.859375,
+      "logits/rejected": -2.78125,
+      "logps/chosen": -145.0,
+      "logps/rejected": -360.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.55859375,
+      "rewards/margins": 25.0,
+      "rewards/rejected": -25.5,
+      "step": 850
+    },
+    {
+      "epoch": 2.747603833865815,
+      "grad_norm": 0.00597818443333345,
+      "learning_rate": 4.674556213017751e-08,
+      "logits/chosen": -2.953125,
+      "logits/rejected": -2.859375,
+      "logps/chosen": -135.0,
+      "logps/rejected": -380.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6484375,
+      "rewards/margins": 26.375,
+      "rewards/rejected": -27.0,
+      "step": 860
+    },
+    {
+      "epoch": 2.779552715654952,
+      "grad_norm": 0.0013466028912008678,
+      "learning_rate": 4.082840236686391e-08,
+      "logits/chosen": -2.765625,
+      "logits/rejected": -2.796875,
+      "logps/chosen": -107.0,
+      "logps/rejected": -366.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.37109375,
+      "rewards/margins": 25.75,
+      "rewards/rejected": -26.125,
+      "step": 870
+    },
+    {
+      "epoch": 2.8115015974440896,
+      "grad_norm": 2.7058882274462487e-05,
+      "learning_rate": 3.4911242603550294e-08,
+      "logits/chosen": -2.78125,
+      "logits/rejected": -2.78125,
+      "logps/chosen": -99.0,
+      "logps/rejected": -376.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1787109375,
+      "rewards/margins": 27.125,
+      "rewards/rejected": -27.25,
+      "step": 880
+    },
+    {
+      "epoch": 2.8434504792332267,
+      "grad_norm": 0.002179423905695273,
+      "learning_rate": 2.8994082840236687e-08,
+      "logits/chosen": -2.84375,
+      "logits/rejected": -2.84375,
+      "logps/chosen": -118.5,
+      "logps/rejected": -376.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.5703125,
+      "rewards/margins": 26.375,
+      "rewards/rejected": -26.875,
+      "step": 890
+    },
+    {
+      "epoch": 2.8753993610223643,
+      "grad_norm": 0.0006890554421005211,
+      "learning_rate": 2.3076923076923076e-08,
+      "logits/chosen": -2.765625,
+      "logits/rejected": -2.75,
+      "logps/chosen": -112.0,
+      "logps/rejected": -364.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.42578125,
+      "rewards/margins": 25.75,
+      "rewards/rejected": -26.25,
+      "step": 900
+    },
+    {
+      "epoch": 2.9073482428115014,
+      "grad_norm": 0.0017781976127544882,
+      "learning_rate": 1.7159763313609465e-08,
+      "logits/chosen": -2.8125,
+      "logits/rejected": -2.84375,
+      "logps/chosen": -121.5,
+      "logps/rejected": -350.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.67578125,
+      "rewards/margins": 23.875,
+      "rewards/rejected": -24.5,
+      "step": 910
+    },
+    {
+      "epoch": 2.939297124600639,
+      "grad_norm": 0.00157882752297583,
+      "learning_rate": 1.1242603550295858e-08,
+      "logits/chosen": -2.765625,
+      "logits/rejected": -2.78125,
+      "logps/chosen": -105.5,
+      "logps/rejected": -388.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.1259765625,
+      "rewards/margins": 28.0,
+      "rewards/rejected": -28.125,
+      "step": 920
+    },
+    {
+      "epoch": 2.9712460063897765,
+      "grad_norm": 0.0018180020601003572,
+      "learning_rate": 5.325443786982248e-09,
+      "logits/chosen": -2.828125,
+      "logits/rejected": -2.8125,
+      "logps/chosen": -139.0,
+      "logps/rejected": -370.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.4609375,
+      "rewards/margins": 25.875,
+      "rewards/rejected": -26.25,
+      "step": 930
+    },
+    {
+      "epoch": 3.0,
+      "eval_logits/chosen": -2.96875,
+      "eval_logits/rejected": -2.90625,
+      "eval_logps/chosen": -173.0,
+      "eval_logps/rejected": -350.0,
+      "eval_loss": 0.004805178381502628,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": -0.515625,
+      "eval_rewards/margins": 23.0,
+      "eval_rewards/rejected": -23.625,
+      "eval_runtime": 16.1771,
+      "eval_samples_per_second": 12.363,
+      "eval_steps_per_second": 0.433,
+      "step": 939
     }
   ],
   "logging_steps": 10,
@@ -998,7 +1479,7 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": false
+        "should_training_stop": true
       },
       "attributes": {}
     }