Training in progress, step 117, checkpoint

last-checkpoint/adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:acba92f70feb375a18e9e2df852cbcd5da182fcb0197d06d0d58f8ae3ddefa59
+oid sha256:8f881f7af17e8e692fe8e56b0bb3efdf3d1b82897cd46e96e09ffe05b97b0277
 size 550593856

last-checkpoint/global_step117/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:b022a2efa834e66981a88b542ee77c9a22d1cc9d3836eab0ad8b25b22bedb412
+size 243590464

last-checkpoint/global_step117/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:ecefa1c476a3924adc0f5af8e5cd57ae04d9ec8b6aa7fa424220296ed6de783a
+size 243590464

last-checkpoint/global_step117/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:ed3b0a80362e19b37f9382e8259885df8f3431851e62c1817798aa18507ecd96
+size 243590464

last-checkpoint/global_step117/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:c1fabbb4467f60e2d2f838e0d7904d5c62e0bebc24587b77b004961b78091a03
+size 243590464

last-checkpoint/global_step117/bf16_zero_pp_rank_4_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cdcd5e95d483c18f1fa2ec0ba3084247f5324a1f3f5e1be3efa33c4b0cd6a152
+size 243590464

last-checkpoint/global_step117/bf16_zero_pp_rank_5_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d56412bbd0b12661c93f11ce1ac2c81bc4bf38436dd305b9d3708fdb58e92836
+size 243590464

last-checkpoint/global_step117/bf16_zero_pp_rank_6_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15a80a16ca24cea74922ea27bb63f4c4ef2b3122a81ea5d6b90114913aed44a3
+size 243590464

last-checkpoint/global_step117/bf16_zero_pp_rank_7_mp_rank_00_optim_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c575316af53d3fae6b90c2e2c15e9fb6a2e444df010ff98b51357cbb4bca039
+size 243590464

last-checkpoint/global_step117/zero_pp_rank_0_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3e65a59c2801adf6a4cf0311b2a04c92f2fb5afe09639152226c1a505442df4
+size 211435686

last-checkpoint/global_step117/zero_pp_rank_1_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:30c6b92167c09ff2301dac0aefd7a9b2b2e1aa0e79f037c85910a2638e0fabd1
+size 211435686

last-checkpoint/global_step117/zero_pp_rank_2_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0649d8e17c76b6fba7134c3130b44ba1f2a4432f77dd6cb4a6f543f76c3c1f59
+size 211435686

last-checkpoint/global_step117/zero_pp_rank_3_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:265ea6d998aa9864e5031343514b3b5e1183aea55e6cfaa61e36002e616bddd7
+size 211435686

last-checkpoint/global_step117/zero_pp_rank_4_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:500d032f2e3ed875432838e8efefe71533914c8532b820c6ecc8bf364e037eff
+size 211435686

last-checkpoint/global_step117/zero_pp_rank_5_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f116d708be0a4523d9ad04955e59ad571408245ce19a62e8e992a1001535f56d
+size 211435686

last-checkpoint/global_step117/zero_pp_rank_6_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a6e0edd58f6ec4f4a55e8a0c5284c950ce0157996ba5ac7ee075f875400d983
+size 211435686

last-checkpoint/global_step117/zero_pp_rank_7_mp_rank_00_model_states.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ef48f47b30be1fef2e9d62aa10c338fb421585678c988fb05ffc1d97d9629bb7
+size 211435686

last-checkpoint/latest

@@ -1 +1 @@
-global_step78
+global_step117

last-checkpoint/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7be6a7458fb39bc70e00e530259e7abd91632050963220c5e5345ba476614f0f
+oid sha256:aac43c4e4aa943b1080df7e7983872659b6a36d1bde144f8865b626eeee7f434
 size 15920

last-checkpoint/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:66026e8691785fbd170c2d9ae99d3724845775d9ad2d10b4043deaa40f54d9d3
+oid sha256:f7c947232f2bdf46ca9ace1c9ebaf1e33912f79599e506f7f7c8ff48a11c7b15
 size 15920

last-checkpoint/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:30d6830c96020be339f5ec273931b81b6fdcd4cef823d11d94c4fd943a776041
+oid sha256:8cfa2ab32c5e9f8e4c49775769b133f4238f016c4e48e8b12b51160d4afbf783
 size 15920

last-checkpoint/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:05af7d968baba27b3aae094ad8d163a6fe1341c5eb4755a791de6267a3ac4f30
+oid sha256:c73c8b459dde07f85f2cffa5de5fb0bf136e4a80afc50d2475bc570d14a65b2e
 size 15920

last-checkpoint/rng_state_4.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6baae28919653bdf03858f680a4661508c0687d95f095cac532ad969406a1502
+oid sha256:d9533672e6284f7cd4c7c4acb5ad621a319e75dbe8a3e90685943824a71eea36
 size 15920

last-checkpoint/rng_state_5.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c83b8d7bb4cbda572628869e88e850db71c0d86c862de24379b95fd3657c08b7
+oid sha256:7a9be2d2ec0fbf037c14fb0f5f0467f7c960e90414356feec416f8eeb1684671
 size 15920

last-checkpoint/rng_state_6.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7dd53716af0f7a80bcb94ac73844bb0733bac101ccb526edad0aec4835119044
+oid sha256:c6a4539a42d09819ea8a238c4dc3e947e079d2a1390a7babff9377b3718e2a09
 size 15920

last-checkpoint/rng_state_7.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:00703c12bd26cde475b80092b041d86d78da55e5a8dff2fa4a8eae5ee025f617
+oid sha256:e097a64585a03993feb0a88a81b38258f3325c6b6c17bdedb7335173ed053ac4
 size 15920

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:263ec1efc1043d4f3446415e08ff68d883fbbe8fdcc5976db1d7ca97cdc780b3
+oid sha256:924db454eec3725afade5f0aae6c3ada154b140ee15da5b50e9bc51987b43065
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.203125,
+  "epoch": 0.3046875,
   "eval_steps": 39,
-  "global_step": 78,
+  "global_step": 117,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -577,6 +577,287 @@
       "eval_samples_per_second": 1.223,
       "eval_steps_per_second": 0.153,
       "step": 78
+    },
+    {
+      "epoch": 0.20572916666666666,
+      "grad_norm": 0.14467889640535875,
+      "learning_rate": 9.42907532874657e-05,
+      "loss": 2.4769,
+      "step": 79
+    },
+    {
+      "epoch": 0.20833333333333334,
+      "grad_norm": 0.13503613492125413,
+      "learning_rate": 9.40999585926253e-05,
+      "loss": 2.4346,
+      "step": 80
+    },
+    {
+      "epoch": 0.2109375,
+      "grad_norm": 0.1517728973900508,
+      "learning_rate": 9.390625136662972e-05,
+      "loss": 2.261,
+      "step": 81
+    },
+    {
+      "epoch": 0.21354166666666666,
+      "grad_norm": 0.1466563383653839,
+      "learning_rate": 9.370964603860753e-05,
+      "loss": 2.5261,
+      "step": 82
+    },
+    {
+      "epoch": 0.21614583333333334,
+      "grad_norm": 0.14618631155344838,
+      "learning_rate": 9.351015725356514e-05,
+      "loss": 2.4269,
+      "step": 83
+    },
+    {
+      "epoch": 0.21875,
+      "grad_norm": 0.15601390292939937,
+      "learning_rate": 9.33077998712958e-05,
+      "loss": 2.3773,
+      "step": 84
+    },
+    {
+      "epoch": 0.22135416666666666,
+      "grad_norm": 0.14415725413526403,
+      "learning_rate": 9.310258896527278e-05,
+      "loss": 2.3473,
+      "step": 85
+    },
+    {
+      "epoch": 0.22395833333333334,
+      "grad_norm": 0.13470412653239092,
+      "learning_rate": 9.289453982152653e-05,
+      "loss": 2.3472,
+      "step": 86
+    },
+    {
+      "epoch": 0.2265625,
+      "grad_norm": 0.15570776289510233,
+      "learning_rate": 9.2683667937506e-05,
+      "loss": 2.4414,
+      "step": 87
+    },
+    {
+      "epoch": 0.22916666666666666,
+      "grad_norm": 0.14612626399487072,
+      "learning_rate": 9.246998902092428e-05,
+      "loss": 2.4655,
+      "step": 88
+    },
+    {
+      "epoch": 0.23177083333333334,
+      "grad_norm": 0.18267625807633459,
+      "learning_rate": 9.22535189885886e-05,
+      "loss": 2.2771,
+      "step": 89
+    },
+    {
+      "epoch": 0.234375,
+      "grad_norm": 0.16065010725957032,
+      "learning_rate": 9.203427396521454e-05,
+      "loss": 2.4194,
+      "step": 90
+    },
+    {
+      "epoch": 0.23697916666666666,
+      "grad_norm": 0.18321820032984756,
+      "learning_rate": 9.181227028222508e-05,
+      "loss": 2.4456,
+      "step": 91
+    },
+    {
+      "epoch": 0.23958333333333334,
+      "grad_norm": 0.13929464732035163,
+      "learning_rate": 9.158752447653397e-05,
+      "loss": 2.3989,
+      "step": 92
+    },
+    {
+      "epoch": 0.2421875,
+      "grad_norm": 0.15585898969003706,
+      "learning_rate": 9.136005328931395e-05,
+      "loss": 2.4777,
+      "step": 93
+    },
+    {
+      "epoch": 0.24479166666666666,
+      "grad_norm": 0.14793073723102618,
+      "learning_rate": 9.112987366474972e-05,
+      "loss": 2.379,
+      "step": 94
+    },
+    {
+      "epoch": 0.24739583333333334,
+      "grad_norm": 0.14075868568691305,
+      "learning_rate": 9.089700274877574e-05,
+      "loss": 2.3281,
+      "step": 95
+    },
+    {
+      "epoch": 0.25,
+      "grad_norm": 0.15327339815586916,
+      "learning_rate": 9.066145788779908e-05,
+      "loss": 2.4517,
+      "step": 96
+    },
+    {
+      "epoch": 0.2526041666666667,
+      "grad_norm": 0.15505433166839452,
+      "learning_rate": 9.042325662740726e-05,
+      "loss": 2.2919,
+      "step": 97
+    },
+    {
+      "epoch": 0.2552083333333333,
+      "grad_norm": 0.1468856057426848,
+      "learning_rate": 9.018241671106134e-05,
+      "loss": 2.3889,
+      "step": 98
+    },
+    {
+      "epoch": 0.2578125,
+      "grad_norm": 0.1478488710337695,
+      "learning_rate": 8.993895607877418e-05,
+      "loss": 2.3595,
+      "step": 99
+    },
+    {
+      "epoch": 0.2604166666666667,
+      "grad_norm": 0.175359520279898,
+      "learning_rate": 8.969289286577408e-05,
+      "loss": 2.553,
+      "step": 100
+    },
+    {
+      "epoch": 0.2630208333333333,
+      "grad_norm": 0.14503182442700818,
+      "learning_rate": 8.9444245401154e-05,
+      "loss": 2.3573,
+      "step": 101
+    },
+    {
+      "epoch": 0.265625,
+      "grad_norm": 0.14177870868430487,
+      "learning_rate": 8.919303220650606e-05,
+      "loss": 2.3274,
+      "step": 102
+    },
+    {
+      "epoch": 0.2682291666666667,
+      "grad_norm": 0.1779007031093982,
+      "learning_rate": 8.893927199454207e-05,
+      "loss": 2.2008,
+      "step": 103
+    },
+    {
+      "epoch": 0.2708333333333333,
+      "grad_norm": 0.14578869126284813,
+      "learning_rate": 8.868298366769954e-05,
+      "loss": 2.444,
+      "step": 104
+    },
+    {
+      "epoch": 0.2734375,
+      "grad_norm": 0.14046914929392473,
+      "learning_rate": 8.842418631673365e-05,
+      "loss": 2.2844,
+      "step": 105
+    },
+    {
+      "epoch": 0.2760416666666667,
+      "grad_norm": 0.13720708755401478,
+      "learning_rate": 8.816289921929516e-05,
+      "loss": 2.5047,
+      "step": 106
+    },
+    {
+      "epoch": 0.2786458333333333,
+      "grad_norm": 0.13901967393067377,
+      "learning_rate": 8.789914183849449e-05,
+      "loss": 2.4602,
+      "step": 107
+    },
+    {
+      "epoch": 0.28125,
+      "grad_norm": 0.14165507438668587,
+      "learning_rate": 8.763293382145195e-05,
+      "loss": 2.3135,
+      "step": 108
+    },
+    {
+      "epoch": 0.2838541666666667,
+      "grad_norm": 0.1588574639328753,
+      "learning_rate": 8.73642949978341e-05,
+      "loss": 2.3731,
+      "step": 109
+    },
+    {
+      "epoch": 0.2864583333333333,
+      "grad_norm": 0.1905241886636867,
+      "learning_rate": 8.709324537837684e-05,
+      "loss": 2.4152,
+      "step": 110
+    },
+    {
+      "epoch": 0.2890625,
+      "grad_norm": 0.16919809232424363,
+      "learning_rate": 8.681980515339464e-05,
+      "loss": 2.2948,
+      "step": 111
+    },
+    {
+      "epoch": 0.2916666666666667,
+      "grad_norm": 0.15114015395920236,
+      "learning_rate": 8.654399469127673e-05,
+      "loss": 2.4309,
+      "step": 112
+    },
+    {
+      "epoch": 0.2942708333333333,
+      "grad_norm": 0.13890547261939354,
+      "learning_rate": 8.626583453696976e-05,
+      "loss": 2.4866,
+      "step": 113
+    },
+    {
+      "epoch": 0.296875,
+      "grad_norm": 0.1490334985258922,
+      "learning_rate": 8.598534541044747e-05,
+      "loss": 2.4748,
+      "step": 114
+    },
+    {
+      "epoch": 0.2994791666666667,
+      "grad_norm": 0.15113023404371403,
+      "learning_rate": 8.570254820516728e-05,
+      "loss": 2.3583,
+      "step": 115
+    },
+    {
+      "epoch": 0.3020833333333333,
+      "grad_norm": 0.14358757342098682,
+      "learning_rate": 8.541746398651395e-05,
+      "loss": 2.4005,
+      "step": 116
+    },
+    {
+      "epoch": 0.3046875,
+      "grad_norm": 0.1494340804949735,
+      "learning_rate": 8.513011399023036e-05,
+      "loss": 2.3249,
+      "step": 117
+    },
+    {
+      "epoch": 0.3046875,
+      "eval_loss": 2.4119415283203125,
+      "eval_runtime": 65.4329,
+      "eval_samples_per_second": 1.223,
+      "eval_steps_per_second": 0.153,
+      "step": 117
     }
   ],
   "logging_steps": 1,
@@ -596,7 +877,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 25785104596992.0,
+  "total_flos": 38677656895488.0,
   "train_batch_size": 1,
   "trial_name": null,
   "trial_params": null