Training in progress, epoch 2, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:27b99e44d1642ff35eac102e28acb8fb9b3bb424f9a0a915f181af67bcc27a99
+oid sha256:94145901cac5d37530f80471e4b4b3bbc14f69bace881edc55e79c4a280b2fe1
 size 457905896

last-checkpoint/global_step1000/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30b50aa9b1603c37097cbc41ec9e2a9df97b89aa5ab5b4c51a03122aca7a0329
+size 345211429

last-checkpoint/global_step1000/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:fc6c74474ac594de3cfccb6deb851c103153f655b12b3726fb7b01214a793b82
+size 345211429

last-checkpoint/global_step1000/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:839554996d9963e71e6826b5728cc129aafd216129db4bf9dcca790fe94a32a8
+size 345211493

last-checkpoint/global_step1000/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:b78c0e444825c963a5ab57579588bebb9ee72cae06678ddec53b3daa25f44605
+size 345211493

last-checkpoint/global_step1000/mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ea7a7089ec9a9cb1a80ef6c513c451cd2509d376bc753cdb760a8cdee0f0dc3a
+size 458029401

last-checkpoint/latest

@@ -1 +1 @@
-global_step950
+global_step1000

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:83ad0934a2d6b3f85e444be8315b90d4cdc2cccd707bd8579a395f8976aadff6
+oid sha256:1be13dd8b4a3bc7b701e88946b29a7b87380d96c425b6d5691a0f69ae3ab3fe5
 size 15429

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:914d285e2670dfd51e775d793877ab7652a127652eccbc31c997add35e0f0d39
+oid sha256:fafde2dd034bc39a2e31703b9cf0035b303aa84d531cdd8155f396e175e8fa4b
 size 15429

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:00fc02891b6172a4a6fa9ae8716dcf4e030bb69e83897009c444a2b4411521cf
+oid sha256:753fffd4deb14ac0f552b1129f7e5cee20389729e2c1e37fa5ec43126284e923
 size 15429

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5ab76950626f09b4be6269d69413423468034295ae98c7402b8b8027c10409f4
+oid sha256:10c1216e6af74878b7df21abdf219031113370d1462f68fb96203d87328e26a0
 size 15429

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e674051f2082998170670e809bb0cb7b30d352a675738a283e03791944312bb5
+oid sha256:7953c0bf1faf7e50d2528974bea59bebbf12324b8bce375b2662393679d16ddf
 size 1401

last-checkpoint/trainer_state.json

@@ -1,10 +1,10 @@
 {
   "best_global_step": null,
-  "best_metric": 0.08488892763853073,
+  "best_metric": 0.08447162061929703,
   "best_model_checkpoint": null,
-  "epoch": 2.1345668629100083,
+  "epoch": 2.2467059153350153,
   "eval_steps": 50,
-  "global_step": 950,
+  "global_step": 1000,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -3162,6 +3162,172 @@
       "eval_samples_per_second": 120.652,
       "eval_steps_per_second": 1.885,
       "step": 950
+    },
+    {
+      "epoch": 2.145780768152509,
+      "grad_norm": 0.1978345364332199,
+      "learning_rate": 0.00011790818448805937,
+      "logits/chosen": -3.3980469703674316,
+      "logits/rejected": -3.207812547683716,
+      "logps/chosen": -612.0999755859375,
+      "logps/rejected": -766.7999877929688,
+      "loss": 0.0966,
+      "rewards/accuracies": 0.967578113079071,
+      "rewards/chosen": -21.609375,
+      "rewards/margins": 16.5859375,
+      "rewards/rejected": -38.20624923706055,
+      "step": 955
+    },
+    {
+      "epoch": 2.1569946733950096,
+      "grad_norm": 0.31681209802627563,
+      "learning_rate": 0.00011788260577323539,
+      "logits/chosen": -3.346484422683716,
+      "logits/rejected": -3.173046827316284,
+      "logps/chosen": -598.5,
+      "logps/rejected": -735.0999755859375,
+      "loss": 0.0909,
+      "rewards/accuracies": 0.9574218988418579,
+      "rewards/chosen": -20.870311737060547,
+      "rewards/margins": 15.092187881469727,
+      "rewards/rejected": -35.990623474121094,
+      "step": 960
+    },
+    {
+      "epoch": 2.1682085786375107,
+      "grad_norm": 0.3784012794494629,
+      "learning_rate": 0.00011785688146466574,
+      "logits/chosen": -3.3734374046325684,
+      "logits/rejected": -3.19921875,
+      "logps/chosen": -611.7000122070312,
+      "logps/rejected": -742.2999877929688,
+      "loss": 0.0931,
+      "rewards/accuracies": 0.9644531011581421,
+      "rewards/chosen": -21.612499237060547,
+      "rewards/margins": 14.442187309265137,
+      "rewards/rejected": -36.04375076293945,
+      "step": 965
+    },
+    {
+      "epoch": 2.1794224838800114,
+      "grad_norm": 0.09592423588037491,
+      "learning_rate": 0.00011783101162711136,
+      "logits/chosen": -3.405078172683716,
+      "logits/rejected": -3.2359375953674316,
+      "logps/chosen": -605.4000244140625,
+      "logps/rejected": -745.2999877929688,
+      "loss": 0.0868,
+      "rewards/accuracies": 0.9605468511581421,
+      "rewards/chosen": -21.037500381469727,
+      "rewards/margins": 15.056249618530273,
+      "rewards/rejected": -36.087501525878906,
+      "step": 970
+    },
+    {
+      "epoch": 2.190636389122512,
+      "grad_norm": 0.330229252576828,
+      "learning_rate": 0.00011780499632569951,
+      "logits/chosen": -3.3851561546325684,
+      "logits/rejected": -3.201171875,
+      "logps/chosen": -625.4000244140625,
+      "logps/rejected": -769.2999877929688,
+      "loss": 0.0989,
+      "rewards/accuracies": 0.95703125,
+      "rewards/chosen": -22.649999618530273,
+      "rewards/margins": 15.623437881469727,
+      "rewards/rejected": -38.256248474121094,
+      "step": 975
+    },
+    {
+      "epoch": 2.2018502943650127,
+      "grad_norm": 0.12297901511192322,
+      "learning_rate": 0.0001177788356259237,
+      "logits/chosen": -3.4078125953674316,
+      "logits/rejected": -3.2144532203674316,
+      "logps/chosen": -641.7999877929688,
+      "logps/rejected": -799.4000244140625,
+      "loss": 0.0808,
+      "rewards/accuracies": 0.969531238079071,
+      "rewards/chosen": -24.024999618530273,
+      "rewards/margins": 16.559375762939453,
+      "rewards/rejected": -40.57500076293945,
+      "step": 980
+    },
+    {
+      "epoch": 2.2130641996075133,
+      "grad_norm": 0.15143366158008575,
+      "learning_rate": 0.00011775252959364346,
+      "logits/chosen": -3.4164061546325684,
+      "logits/rejected": -3.2125000953674316,
+      "logps/chosen": -671.2000122070312,
+      "logps/rejected": -818.0999755859375,
+      "loss": 0.0833,
+      "rewards/accuracies": 0.9605468511581421,
+      "rewards/chosen": -26.240625381469727,
+      "rewards/margins": 16.1875,
+      "rewards/rejected": -42.443748474121094,
+      "step": 985
+    },
+    {
+      "epoch": 2.224278104850014,
+      "grad_norm": 0.17936468124389648,
+      "learning_rate": 0.00011772607829508418,
+      "logits/chosen": -3.48828125,
+      "logits/rejected": -3.3042969703674316,
+      "logps/chosen": -682.5999755859375,
+      "logps/rejected": -827.0999755859375,
+      "loss": 0.0892,
+      "rewards/accuracies": 0.9609375,
+      "rewards/chosen": -27.346874237060547,
+      "rewards/margins": 15.542187690734863,
+      "rewards/rejected": -42.89374923706055,
+      "step": 990
+    },
+    {
+      "epoch": 2.2354920100925146,
+      "grad_norm": 0.2282583862543106,
+      "learning_rate": 0.00011769948179683699,
+      "logits/chosen": -3.513671875,
+      "logits/rejected": -3.329296827316284,
+      "logps/chosen": -664.5,
+      "logps/rejected": -816.0999755859375,
+      "loss": 0.0901,
+      "rewards/accuracies": 0.961718738079071,
+      "rewards/chosen": -25.475000381469727,
+      "rewards/margins": 16.407812118530273,
+      "rewards/rejected": -41.896873474121094,
+      "step": 995
+    },
+    {
+      "epoch": 2.2467059153350153,
+      "grad_norm": 0.19659578800201416,
+      "learning_rate": 0.00011767274016585854,
+      "logits/chosen": -3.522656202316284,
+      "logits/rejected": -3.35546875,
+      "logps/chosen": -633.0999755859375,
+      "logps/rejected": -791.0999755859375,
+      "loss": 0.0839,
+      "rewards/accuracies": 0.966796875,
+      "rewards/chosen": -23.706249237060547,
+      "rewards/margins": 16.396875381469727,
+      "rewards/rejected": -40.09375,
+      "step": 1000
+    },
+    {
+      "epoch": 2.2467059153350153,
+      "eval_logits/chosen": -3.5166471004486084,
+      "eval_logits/rejected": -3.3513433933258057,
+      "eval_logps/chosen": -630.168212890625,
+      "eval_logps/rejected": -778.5046997070312,
+      "eval_loss": 0.08447162061929703,
+      "eval_rewards/accuracies": 0.9664135575294495,
+      "eval_rewards/chosen": -23.545560836791992,
+      "eval_rewards/margins": 15.758177757263184,
+      "eval_rewards/rejected": -39.31191635131836,
+      "eval_runtime": 56.7078,
+      "eval_samples_per_second": 120.759,
+      "eval_steps_per_second": 1.887,
+      "step": 1000
     }
   ],
   "logging_steps": 5,