Training in progress, epoch 3, checkpoint

last-checkpoint/global_step396/bf16_zero_pp_rank_0_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:75fec04bb7ecc38e12ed2e5a1a3241993e65368f3129503bfc1a52fe1af844af
+size 13476836524

last-checkpoint/global_step396/bf16_zero_pp_rank_1_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:063e0745190972883fb0712e4d901288ac19869d470854e81073e3b9a24075fc
+size 13476836524

last-checkpoint/global_step396/bf16_zero_pp_rank_2_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:035fbdde4dc629a97279f445bfa7b063e3a447b548d63b89cbdb2722fa1e9c09
+size 13476836524

last-checkpoint/global_step396/bf16_zero_pp_rank_3_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:89f31ddd110bfb53e3f168b00c3a9de2e308f6647d76f73d19d21550fabd88ee
+size 13476836524

last-checkpoint/global_step396/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5962b2aca10696a8628b10cbce7ad78a11000b3fac1d95a756963951ceffad3e
+size 150693

last-checkpoint/global_step396/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3691154be3e3e35e1d4e0345e8aeacdb622d3637e3091010ddb0619d9d658649
+size 150693

last-checkpoint/global_step396/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4d4e16b1c445fd3010969c2e4ce5b31c7d615bd0e015a33c0eeb08d188183b55
+size 150693

last-checkpoint/global_step396/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6722ba9b3365e1978cc112963745f06b86d81e638b5bbd3b50600f968f5630be
+size 150693

last-checkpoint/latest

@@ -1 +1 @@
-global_step132
+global_step396

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:37e92413357e82161c64e7455ccc831c04a3f7526efd54caee0edf972f72b2ef
+oid sha256:05d5679587549d4fad07eeaf2ad1483b371699533a9a506f35fcd0fb42d4766a
 size 4938985352

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f4f2dee00fac343d1d586af2fa0bb6776f5bc80b9482cbc412f71755841d7e36
+oid sha256:3e60d2612906bc653ce6f4e818aebb9c802ae3879c90111a3263e0bb8b8a0455
 size 4947390880

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

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:b9b795516c93d3cb07fbd20f45b3aa52076fcec87e6498f2eccd4bc512117c1b
+oid sha256:1c580cd2e18577ae23e6e5bd6c9b19c020082256c9f8c280cf6523edec4414b5
 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:ab7cf0d34d60dfda516cf9661904550e2e294e723edd07c25c738f05e8ba92d1
 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:f7b183324e8227a51a9556d86b2ad893a8c4c52205ed4a737356c6611dac7353
 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:ac69e994090f4818cb1fa6f6cefa363178552c3c731c6507ff195bcb07fd5bef
 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:f68e61b63402f8afb1f69c960f7944965655dac11e3ccf29919c282f23931f86
 size 15024

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:081681bbbe98e112e0df8f351c9ba3661c598d1ed2ac20b20f26378a5328c1e6
+oid sha256:bd42a1a947062e05f4c7c69eaa1bb1d4fd05329e7438195743e9f0a70c1343ae
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.0,
+  "epoch": 3.0,
   "eval_steps": 500,
-  "global_step": 132,
+  "global_step": 396,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -233,6 +233,428 @@
       "eval_samples_per_second": 16.357,
       "eval_steps_per_second": 0.584,
       "step": 132
+    },
+    {
+      "epoch": 1.0606060606060606,
+      "grad_norm": 0.044137608822413354,
+      "learning_rate": 3.5955056179775277e-07,
+      "logits/chosen": -1.578125,
+      "logits/rejected": -1.5078125,
+      "logps/chosen": -155.0,
+      "logps/rejected": -260.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.65625,
+      "rewards/margins": 11.3125,
+      "rewards/rejected": -12.0,
+      "step": 140
+    },
+    {
+      "epoch": 1.1363636363636362,
+      "grad_norm": 0.03726527729430989,
+      "learning_rate": 3.4550561797752807e-07,
+      "logits/chosen": -1.578125,
+      "logits/rejected": -1.515625,
+      "logps/chosen": -152.0,
+      "logps/rejected": -252.0,
+      "loss": 0.0017,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.427734375,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.4375,
+      "step": 150
+    },
+    {
+      "epoch": 1.2121212121212122,
+      "grad_norm": 0.04249817642262755,
+      "learning_rate": 3.314606741573033e-07,
+      "logits/chosen": -1.5625,
+      "logits/rejected": -1.5703125,
+      "logps/chosen": -157.0,
+      "logps/rejected": -252.0,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.345703125,
+      "rewards/margins": 10.9375,
+      "rewards/rejected": -11.25,
+      "step": 160
+    },
+    {
+      "epoch": 1.2878787878787878,
+      "grad_norm": 0.02636991212367461,
+      "learning_rate": 3.1741573033707863e-07,
+      "logits/chosen": -1.546875,
+      "logits/rejected": -1.5546875,
+      "logps/chosen": -153.0,
+      "logps/rejected": -253.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.326171875,
+      "rewards/margins": 11.0,
+      "rewards/rejected": -11.3125,
+      "step": 170
+    },
+    {
+      "epoch": 1.3636363636363638,
+      "grad_norm": 0.002485552309390973,
+      "learning_rate": 3.0337078651685393e-07,
+      "logits/chosen": -1.59375,
+      "logits/rejected": -1.578125,
+      "logps/chosen": -170.0,
+      "logps/rejected": -251.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.4375,
+      "rewards/margins": 11.0625,
+      "rewards/rejected": -11.5,
+      "step": 180
+    },
+    {
+      "epoch": 1.4393939393939394,
+      "grad_norm": 0.06372981794712247,
+      "learning_rate": 2.893258426966292e-07,
+      "logits/chosen": -1.546875,
+      "logits/rejected": -1.53125,
+      "logps/chosen": -152.0,
+      "logps/rejected": -254.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.384765625,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.0625,
+      "step": 190
+    },
+    {
+      "epoch": 1.5151515151515151,
+      "grad_norm": 0.056568516864774675,
+      "learning_rate": 2.752808988764045e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.5078125,
+      "logps/chosen": -158.0,
+      "logps/rejected": -256.0,
+      "loss": 0.0011,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.53515625,
+      "rewards/margins": 11.5,
+      "rewards/rejected": -12.0,
+      "step": 200
+    },
+    {
+      "epoch": 1.5909090909090908,
+      "grad_norm": 0.0017135050317799595,
+      "learning_rate": 2.612359550561798e-07,
+      "logits/chosen": -1.609375,
+      "logits/rejected": -1.53125,
+      "logps/chosen": -171.0,
+      "logps/rejected": -264.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.6875,
+      "rewards/margins": 11.875,
+      "rewards/rejected": -12.5,
+      "step": 210
+    },
+    {
+      "epoch": 1.6666666666666665,
+      "grad_norm": 0.011438568560823217,
+      "learning_rate": 2.4719101123595505e-07,
+      "logits/chosen": -1.5,
+      "logits/rejected": -1.53125,
+      "logps/chosen": -155.0,
+      "logps/rejected": -264.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.8125,
+      "rewards/margins": 11.6875,
+      "rewards/rejected": -12.5,
+      "step": 220
+    },
+    {
+      "epoch": 1.7424242424242424,
+      "grad_norm": 0.00196532311899783,
+      "learning_rate": 2.331460674157303e-07,
+      "logits/chosen": -1.59375,
+      "logits/rejected": -1.5625,
+      "logps/chosen": -160.0,
+      "logps/rejected": -264.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.98828125,
+      "rewards/margins": 11.75,
+      "rewards/rejected": -12.75,
+      "step": 230
+    },
+    {
+      "epoch": 1.8181818181818183,
+      "grad_norm": 0.02412516870728502,
+      "learning_rate": 2.191011235955056e-07,
+      "logits/chosen": -1.578125,
+      "logits/rejected": -1.578125,
+      "logps/chosen": -163.0,
+      "logps/rejected": -272.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.859375,
+      "rewards/margins": 12.0625,
+      "rewards/rejected": -12.9375,
+      "step": 240
+    },
+    {
+      "epoch": 1.893939393939394,
+      "grad_norm": 0.008677592300417748,
+      "learning_rate": 2.0505617977528089e-07,
+      "logits/chosen": -1.5625,
+      "logits/rejected": -1.546875,
+      "logps/chosen": -164.0,
+      "logps/rejected": -268.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.98046875,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -13.0,
+      "step": 250
+    },
+    {
+      "epoch": 1.9696969696969697,
+      "grad_norm": 0.005504070078399264,
+      "learning_rate": 1.9101123595505617e-07,
+      "logits/chosen": -1.53125,
+      "logits/rejected": -1.515625,
+      "logps/chosen": -164.0,
+      "logps/rejected": -268.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.91796875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.0625,
+      "step": 260
+    },
+    {
+      "epoch": 2.0,
+      "eval_logits/chosen": -1.4921875,
+      "eval_logits/rejected": -1.484375,
+      "eval_logps/chosen": -119.5,
+      "eval_logps/rejected": -256.0,
+      "eval_loss": 0.0008152184309437871,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": -0.79296875,
+      "eval_rewards/margins": 11.375,
+      "eval_rewards/rejected": -12.125,
+      "eval_runtime": 5.7051,
+      "eval_samples_per_second": 14.724,
+      "eval_steps_per_second": 0.526,
+      "step": 264
+    },
+    {
+      "epoch": 2.0454545454545454,
+      "grad_norm": 0.00984660780499755,
+      "learning_rate": 1.7696629213483144e-07,
+      "logits/chosen": -1.5234375,
+      "logits/rejected": -1.5390625,
+      "logps/chosen": -148.0,
+      "logps/rejected": -266.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7734375,
+      "rewards/margins": 12.0,
+      "rewards/rejected": -12.75,
+      "step": 270
+    },
+    {
+      "epoch": 2.121212121212121,
+      "grad_norm": 0.004414538457258972,
+      "learning_rate": 1.6292134831460675e-07,
+      "logits/chosen": -1.5,
+      "logits/rejected": -1.515625,
+      "logps/chosen": -165.0,
+      "logps/rejected": -264.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75390625,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.875,
+      "step": 280
+    },
+    {
+      "epoch": 2.196969696969697,
+      "grad_norm": 0.016662720665351936,
+      "learning_rate": 1.4887640449438203e-07,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.546875,
+      "logps/chosen": -156.0,
+      "logps/rejected": -272.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.79296875,
+      "rewards/margins": 12.1875,
+      "rewards/rejected": -13.0,
+      "step": 290
+    },
+    {
+      "epoch": 2.2727272727272725,
+      "grad_norm": 0.05873531330460493,
+      "learning_rate": 1.3483146067415728e-07,
+      "logits/chosen": -1.59375,
+      "logits/rejected": -1.546875,
+      "logps/chosen": -160.0,
+      "logps/rejected": -268.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.78515625,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.0,
+      "step": 300
+    },
+    {
+      "epoch": 2.3484848484848486,
+      "grad_norm": 0.046015511031248915,
+      "learning_rate": 1.2078651685393259e-07,
+      "logits/chosen": -1.5625,
+      "logits/rejected": -1.546875,
+      "logps/chosen": -151.0,
+      "logps/rejected": -264.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.70703125,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.875,
+      "step": 310
+    },
+    {
+      "epoch": 2.4242424242424243,
+      "grad_norm": 0.3021671776155001,
+      "learning_rate": 1.0674157303370785e-07,
+      "logits/chosen": -1.578125,
+      "logits/rejected": -1.578125,
+      "logps/chosen": -161.0,
+      "logps/rejected": -272.0,
+      "loss": 0.0003,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.75390625,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -13.0625,
+      "step": 320
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.0015387455528242642,
+      "learning_rate": 9.269662921348314e-08,
+      "logits/chosen": -1.6171875,
+      "logits/rejected": -1.5703125,
+      "logps/chosen": -163.0,
+      "logps/rejected": -266.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.74609375,
+      "rewards/margins": 12.375,
+      "rewards/rejected": -13.125,
+      "step": 330
+    },
+    {
+      "epoch": 2.5757575757575757,
+      "grad_norm": 0.001321971934515845,
+      "learning_rate": 7.865168539325842e-08,
+      "logits/chosen": -1.53125,
+      "logits/rejected": -1.609375,
+      "logps/chosen": -150.0,
+      "logps/rejected": -276.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7421875,
+      "rewards/margins": 12.625,
+      "rewards/rejected": -13.375,
+      "step": 340
+    },
+    {
+      "epoch": 2.6515151515151514,
+      "grad_norm": 0.0057061429256677594,
+      "learning_rate": 6.460674157303371e-08,
+      "logits/chosen": -1.53125,
+      "logits/rejected": -1.5703125,
+      "logps/chosen": -159.0,
+      "logps/rejected": -276.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.875,
+      "rewards/margins": 12.5625,
+      "rewards/rejected": -13.4375,
+      "step": 350
+    },
+    {
+      "epoch": 2.7272727272727275,
+      "grad_norm": 0.4762341538722265,
+      "learning_rate": 5.056179775280899e-08,
+      "logits/chosen": -1.5625,
+      "logits/rejected": -1.546875,
+      "logps/chosen": -143.0,
+      "logps/rejected": -268.0,
+      "loss": 0.0004,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.68359375,
+      "rewards/margins": 12.3125,
+      "rewards/rejected": -13.0,
+      "step": 360
+    },
+    {
+      "epoch": 2.8030303030303028,
+      "grad_norm": 0.2900268557706346,
+      "learning_rate": 3.6516853932584266e-08,
+      "logits/chosen": -1.53125,
+      "logits/rejected": -1.5859375,
+      "logps/chosen": -147.0,
+      "logps/rejected": -270.0,
+      "loss": 0.0002,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.85546875,
+      "rewards/margins": 12.25,
+      "rewards/rejected": -13.0625,
+      "step": 370
+    },
+    {
+      "epoch": 2.878787878787879,
+      "grad_norm": 0.005709415780533644,
+      "learning_rate": 2.2471910112359548e-08,
+      "logits/chosen": -1.59375,
+      "logits/rejected": -1.5390625,
+      "logps/chosen": -163.0,
+      "logps/rejected": -270.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.74609375,
+      "rewards/margins": 12.4375,
+      "rewards/rejected": -13.1875,
+      "step": 380
+    },
+    {
+      "epoch": 2.9545454545454546,
+      "grad_norm": 0.009508653721430635,
+      "learning_rate": 8.42696629213483e-09,
+      "logits/chosen": -1.5703125,
+      "logits/rejected": -1.59375,
+      "logps/chosen": -160.0,
+      "logps/rejected": -266.0,
+      "loss": 0.0001,
+      "rewards/accuracies": 1.0,
+      "rewards/chosen": -0.7421875,
+      "rewards/margins": 12.125,
+      "rewards/rejected": -12.875,
+      "step": 390
+    },
+    {
+      "epoch": 3.0,
+      "eval_logits/chosen": -1.5,
+      "eval_logits/rejected": -1.484375,
+      "eval_logps/chosen": -119.0,
+      "eval_logps/rejected": -260.0,
+      "eval_loss": 0.0003405184834264219,
+      "eval_rewards/accuracies": 1.0,
+      "eval_rewards/chosen": -0.734375,
+      "eval_rewards/margins": 11.75,
+      "eval_rewards/rejected": -12.5,
+      "eval_runtime": 16.5049,
+      "eval_samples_per_second": 5.089,
+      "eval_steps_per_second": 0.182,
+      "step": 396
     }
   ],
   "logging_steps": 10,
@@ -247,7 +669,7 @@
         "should_evaluate": false,
         "should_log": false,
         "should_save": true,
-        "should_training_stop": false
+        "should_training_stop": true
       },
       "attributes": {}
     }