Training in progress, epoch 3, checkpoint

last-checkpoint/global_step915/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:e8c3fa2b36b1415c355598292f0e58d07fc38b746a3783a94ffaf12d9d9c69f8
+size 17075366956

last-checkpoint/global_step915/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:b9f61d688d5ffeb48ab98c9c28b0f7f08915d0829cc2fd1579e58f320ca9466c
+size 17075366956

last-checkpoint/global_step915/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:07ed2a5bfebd11a10b97360df8c3cb76d2d53e184381ff46ce90781b793aa594
+size 17075366956

last-checkpoint/global_step915/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:d5c491fc8eea51decda17837070dfe2fde7596e0a04f542030437e928d6cd9e9
+size 17075366956

last-checkpoint/global_step915/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:20a7fd50be01694e49f9de3f8f1272257c8c03f4ad98ad21333c50f043c9f4e2
+size 131892

last-checkpoint/global_step915/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bdc85537b53b2e8316afc05428c47945418ef5624b9e5a3e7fb0883dd92beafa
+size 131828

last-checkpoint/global_step915/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a36623a59edfd11a4af114aed7acad913054b7c916187ceb13f66f370b5df031
+size 131828

last-checkpoint/global_step915/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a971b41276ea7e04ae399c49141ed5b590cd146d7caaaf0c73534af09b8a01c1
+size 131828

last-checkpoint/latest

@@ -1 +1 @@
-global_step610
+global_step915

last-checkpoint/model-00001-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a12c862d2ef96968d2f0f5e4972cea773cd7db53ef6ee1042bcd403afd8b6cd6
+oid sha256:46ee7aa453e70d10c2d280c1ef35378db840e022726ccec385cbea893281f8f4
 size 4995496656

last-checkpoint/model-00002-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c672d54f9e6ba1468c38e46dd54bb25368ce833977ea89a029fa0ff0830a3a7e
+oid sha256:32040c4cc65ef27258a1d1054a03cecd7e32cf87cbfb2f7e1ec1d98dfe5455ce
 size 4982953168

last-checkpoint/model-00003-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c9b6b3c0e7322426b4abba14239ff48c798c2c9ee84e366d2619ac8eab0580c6
+oid sha256:cf9cd027b02fef79cbb1cd7640a0015d6b62b09c3b174cd3ff9f25758d133911
 size 4982953200

last-checkpoint/model-00004-of-00004.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e7b16af01a2e16e531a32351ccf6ff3b7ad91f13d7efe31f409eaa7f8f3a62d6
+oid sha256:d8e5dbff8fcd53770be3cfcdf0dfa0103de5978baec282205d3a2fd64c6fb474
 size 2113988336

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 sha256:23cddb35265adde8fcd130442bb80eda4e7e58a5537171741839d315438de751
+oid sha256:89cc83db19eba05de12a39437c27831b7029da4dd2494fc1e52d31a689e731bb
 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": 610,
+  "global_step": 915,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -969,6 +969,472 @@
       "eval_samples_per_second": 12.919,
       "eval_steps_per_second": 0.461,
       "step": 610
+    },
+    {
+      "epoch": 2.0327868852459017,
+      "grad_norm": 1.7713226584924222e-05,
+      "learning_rate": 1.792223572296476e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -620.0,
+      "logps/rejected": -1240.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.5,
+      "rewards/margins": 53.75,
+      "rewards/rejected": -37.25,
+      "step": 620
+    },
+    {
+      "epoch": 2.0655737704918034,
+      "grad_norm": 1.0249551855925774e-06,
+      "learning_rate": 1.7314702308626974e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -616.0,
+      "logps/rejected": -1176.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.25,
+      "rewards/margins": 52.25,
+      "rewards/rejected": -36.0,
+      "step": 630
+    },
+    {
+      "epoch": 2.098360655737705,
+      "grad_norm": 4.2707376166338526e-05,
+      "learning_rate": 1.6707168894289186e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -656.0,
+      "logps/rejected": -1192.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.625,
+      "rewards/margins": 53.0,
+      "rewards/rejected": -36.5,
+      "step": 640
+    },
+    {
+      "epoch": 2.1311475409836067,
+      "grad_norm": 0.00034207508221282683,
+      "learning_rate": 1.6099635479951396e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -544.0,
+      "logps/rejected": -1240.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 13.8125,
+      "rewards/margins": 51.5,
+      "rewards/rejected": -37.75,
+      "step": 650
+    },
+    {
+      "epoch": 2.1639344262295084,
+      "grad_norm": 5.419595774676902e-09,
+      "learning_rate": 1.5492102065613608e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -482.0,
+      "logps/rejected": -1200.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 12.1875,
+      "rewards/margins": 49.0,
+      "rewards/rejected": -37.0,
+      "step": 660
+    },
+    {
+      "epoch": 2.19672131147541,
+      "grad_norm": 4.851068352220634e-11,
+      "learning_rate": 1.488456865127582e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -752.0,
+      "logps/rejected": -1288.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 19.0,
+      "rewards/margins": 58.5,
+      "rewards/rejected": -39.25,
+      "step": 670
+    },
+    {
+      "epoch": 2.2295081967213113,
+      "grad_norm": 1.7612628984927503e-10,
+      "learning_rate": 1.427703523693803e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -576.0,
+      "logps/rejected": -1192.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 15.5625,
+      "rewards/margins": 53.0,
+      "rewards/rejected": -37.25,
+      "step": 680
+    },
+    {
+      "epoch": 2.262295081967213,
+      "grad_norm": 3.275622407668139e-08,
+      "learning_rate": 1.3669501822600243e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -612.0,
+      "logps/rejected": -1264.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.125,
+      "rewards/margins": 54.75,
+      "rewards/rejected": -38.5,
+      "step": 690
+    },
+    {
+      "epoch": 2.2950819672131146,
+      "grad_norm": 0.0013215912863667734,
+      "learning_rate": 1.3061968408262452e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -580.0,
+      "logps/rejected": -1232.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 14.25,
+      "rewards/margins": 51.0,
+      "rewards/rejected": -36.75,
+      "step": 700
+    },
+    {
+      "epoch": 2.3278688524590163,
+      "grad_norm": 0.00041899600497659085,
+      "learning_rate": 1.2454434993924665e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -516.0,
+      "logps/rejected": -1256.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 13.625,
+      "rewards/margins": 52.5,
+      "rewards/rejected": -38.75,
+      "step": 710
+    },
+    {
+      "epoch": 2.360655737704918,
+      "grad_norm": 7.351246552401219e-09,
+      "learning_rate": 1.1846901579586877e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -584.0,
+      "logps/rejected": -1208.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 14.5,
+      "rewards/margins": 51.5,
+      "rewards/rejected": -37.0,
+      "step": 720
+    },
+    {
+      "epoch": 2.3934426229508197,
+      "grad_norm": 1.2823458114586286e-12,
+      "learning_rate": 1.1239368165249088e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -628.0,
+      "logps/rejected": -1256.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 14.8125,
+      "rewards/margins": 53.5,
+      "rewards/rejected": -38.75,
+      "step": 730
+    },
+    {
+      "epoch": 2.4262295081967213,
+      "grad_norm": 2.2183065144506676e-05,
+      "learning_rate": 1.0631834750911299e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -672.0,
+      "logps/rejected": -1272.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 15.25,
+      "rewards/margins": 54.5,
+      "rewards/rejected": -39.0,
+      "step": 740
+    },
+    {
+      "epoch": 2.459016393442623,
+      "grad_norm": 9.496741561479611e-10,
+      "learning_rate": 1.0024301336573512e-07,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -540.0,
+      "logps/rejected": -1216.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 13.4375,
+      "rewards/margins": 52.0,
+      "rewards/rejected": -38.5,
+      "step": 750
+    },
+    {
+      "epoch": 2.4918032786885247,
+      "grad_norm": 6.428681857091136e-08,
+      "learning_rate": 9.416767922235723e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -676.0,
+      "logps/rejected": -1232.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 17.0,
+      "rewards/margins": 54.5,
+      "rewards/rejected": -37.75,
+      "step": 760
+    },
+    {
+      "epoch": 2.5245901639344264,
+      "grad_norm": 2.0982264237925454e-08,
+      "learning_rate": 8.809234507897934e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -672.0,
+      "logps/rejected": -1184.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.25,
+      "rewards/margins": 53.25,
+      "rewards/rejected": -37.0,
+      "step": 770
+    },
+    {
+      "epoch": 2.557377049180328,
+      "grad_norm": 9.460545250549385e-16,
+      "learning_rate": 8.201701093560146e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -676.0,
+      "logps/rejected": -1208.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 14.9375,
+      "rewards/margins": 52.5,
+      "rewards/rejected": -37.75,
+      "step": 780
+    },
+    {
+      "epoch": 2.5901639344262293,
+      "grad_norm": 4.793838989333602e-09,
+      "learning_rate": 7.594167679222357e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -556.0,
+      "logps/rejected": -1272.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 14.25,
+      "rewards/margins": 53.75,
+      "rewards/rejected": -39.5,
+      "step": 790
+    },
+    {
+      "epoch": 2.6229508196721314,
+      "grad_norm": 6.169679361669359e-11,
+      "learning_rate": 6.986634264884568e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -612.0,
+      "logps/rejected": -1200.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 15.3125,
+      "rewards/margins": 53.0,
+      "rewards/rejected": -37.75,
+      "step": 800
+    },
+    {
+      "epoch": 2.6557377049180326,
+      "grad_norm": 8.064196022250346e-08,
+      "learning_rate": 6.37910085054678e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -656.0,
+      "logps/rejected": -1216.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.875,
+      "rewards/margins": 54.75,
+      "rewards/rejected": -37.75,
+      "step": 810
+    },
+    {
+      "epoch": 2.6885245901639343,
+      "grad_norm": 1.1139956025424588e-07,
+      "learning_rate": 5.771567436208991e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -580.0,
+      "logps/rejected": -1224.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 15.3125,
+      "rewards/margins": 52.75,
+      "rewards/rejected": -37.5,
+      "step": 820
+    },
+    {
+      "epoch": 2.721311475409836,
+      "grad_norm": 1.8351283154166078e-11,
+      "learning_rate": 5.164034021871203e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -648.0,
+      "logps/rejected": -1216.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 15.25,
+      "rewards/margins": 52.5,
+      "rewards/rejected": -37.25,
+      "step": 830
+    },
+    {
+      "epoch": 2.7540983606557377,
+      "grad_norm": 5.336889215608692e-08,
+      "learning_rate": 4.5565006075334144e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -664.0,
+      "logps/rejected": -1240.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.25,
+      "rewards/margins": 54.75,
+      "rewards/rejected": -38.5,
+      "step": 840
+    },
+    {
+      "epoch": 2.7868852459016393,
+      "grad_norm": 1.3681049107441818e-10,
+      "learning_rate": 3.9489671931956255e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -680.0,
+      "logps/rejected": -1240.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 16.125,
+      "rewards/margins": 55.25,
+      "rewards/rejected": -39.25,
+      "step": 850
+    },
+    {
+      "epoch": 2.819672131147541,
+      "grad_norm": 2.7039086242151e-05,
+      "learning_rate": 3.341433778857837e-08,
+      "logits/chosen": 274.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -512.0,
+      "logps/rejected": -1240.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 13.0625,
+      "rewards/margins": 51.5,
+      "rewards/rejected": -38.5,
+      "step": 860
+    },
+    {
+      "epoch": 2.8524590163934427,
+      "grad_norm": 4.1661882432785415e-06,
+      "learning_rate": 2.7339003645200486e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 280.0,
+      "logps/chosen": -588.0,
+      "logps/rejected": -1240.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 13.5,
+      "rewards/margins": 52.0,
+      "rewards/rejected": -38.5,
+      "step": 870
+    },
+    {
+      "epoch": 2.8852459016393444,
+      "grad_norm": 4.794762374205389e-10,
+      "learning_rate": 2.12636695018226e-08,
+      "logits/chosen": 274.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -608.0,
+      "logps/rejected": -1208.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 15.3125,
+      "rewards/margins": 53.0,
+      "rewards/rejected": -37.75,
+      "step": 880
+    },
+    {
+      "epoch": 2.918032786885246,
+      "grad_norm": 1.9558841209760312e-05,
+      "learning_rate": 1.5188335358444714e-08,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -560.0,
+      "logps/rejected": -1280.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 14.0,
+      "rewards/margins": 54.0,
+      "rewards/rejected": -40.0,
+      "step": 890
+    },
+    {
+      "epoch": 2.9508196721311473,
+      "grad_norm": 0.00011639924338885354,
+      "learning_rate": 9.113001215066828e-09,
+      "logits/chosen": 274.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -492.0,
+      "logps/rejected": -1232.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 13.375,
+      "rewards/margins": 52.0,
+      "rewards/rejected": -38.5,
+      "step": 900
+    },
+    {
+      "epoch": 2.9836065573770494,
+      "grad_norm": 9.409851507715962e-11,
+      "learning_rate": 3.0376670716889426e-09,
+      "logits/chosen": 276.0,
+      "logits/rejected": 282.0,
+      "logps/chosen": -536.0,
+      "logps/rejected": -1296.0,
+      "loss": 0.0,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": 13.375,
+      "rewards/margins": 54.0,
+      "rewards/rejected": -40.75,
+      "step": 910
+    },
+    {
+      "epoch": 3.0,
+      "eval_logits/chosen": 276.0,
+      "eval_logits/rejected": 282.0,
+      "eval_logps/chosen": -676.0,
+      "eval_logps/rejected": -1264.0,
+      "eval_loss": 1.2766672341513186e-07,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": 17.5,
+      "eval_rewards/margins": 56.5,
+      "eval_rewards/rejected": -39.0,
+      "eval_runtime": 17.7894,
+      "eval_samples_per_second": 11.018,
+      "eval_steps_per_second": 0.393,
+      "step": 915
     }
   ],
   "logging_steps": 10,
@@ -983,7 +1449,7 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": false
+        "should_training_stop": true
       },
       "attributes": {}
     }