Training in progress, epoch 9, checkpoint

last-checkpoint/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1a2175ec71e87f793ff0f1b56c0e4d67ecb39ba87b782cb059bb94ad285c2f53
+oid sha256:dfd0dbfd5a2e202c41c4474c701ce317f0b353244fedc30fcf9035ad3f0cfb4b
 size 2695611744

last-checkpoint/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:042e4a5ad63d06a910df2e74958ddf6def3d06490a31ec05e030713646e66aae
+oid sha256:e71f4096ee32c8679f4882ef8cbd3add01377b6277f82c373625a3635016054e
 size 26261260

last-checkpoint/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cd92451b020a5e32fd17fed687a0bd65992ebcaaa07504fb89486fb3090b6e0b
+oid sha256:ada9c6a6e42dd9a052cc6c90ea555e554cf2266902dc66003e3529ef94dd9d78
 size 15006

last-checkpoint/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:98b9a9b66ad7c2c92d26ff031cce36ce8f9c2ebdb6be883784d837b0e0137568
+oid sha256:17d282d44e6a12e9ca82ec6b42e7f212659678f5fe8c9a58e24a30599040d3e5
 size 1064

last-checkpoint/trainer_state.json

@@ -1,9 +1,9 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 8.0,
+  "epoch": 9.0,
   "eval_steps": 10,
-  "global_step": 1032,
+  "global_step": 1161,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
@@ -1449,6 +1449,188 @@
       "eval_samples_per_second": 22.032,
       "eval_steps_per_second": 5.508,
       "step": 1030
+    },
+    {
+      "epoch": 8.062469497315764,
+      "grad_norm": 306364.5,
+      "learning_rate": 1.8750000000000002e-05,
+      "loss": 458.2917,
+      "step": 1040
+    },
+    {
+      "epoch": 8.062469497315764,
+      "eval_runtime": 19.626,
+      "eval_samples_per_second": 22.012,
+      "eval_steps_per_second": 5.503,
+      "step": 1040
+    },
+    {
+      "epoch": 8.140556368960468,
+      "grad_norm": 287573.96875,
+      "learning_rate": 1.796875e-05,
+      "loss": 459.69,
+      "step": 1050
+    },
+    {
+      "epoch": 8.140556368960468,
+      "eval_runtime": 19.64,
+      "eval_samples_per_second": 21.996,
+      "eval_steps_per_second": 5.499,
+      "step": 1050
+    },
+    {
+      "epoch": 8.218643240605173,
+      "grad_norm": 87142.5625,
+      "learning_rate": 1.71875e-05,
+      "loss": 380.1467,
+      "step": 1060
+    },
+    {
+      "epoch": 8.218643240605173,
+      "eval_runtime": 19.6334,
+      "eval_samples_per_second": 22.003,
+      "eval_steps_per_second": 5.501,
+      "step": 1060
+    },
+    {
+      "epoch": 8.296730112249879,
+      "grad_norm": 301649.96875,
+      "learning_rate": 1.6406250000000002e-05,
+      "loss": 384.5057,
+      "step": 1070
+    },
+    {
+      "epoch": 8.296730112249879,
+      "eval_runtime": 19.6212,
+      "eval_samples_per_second": 22.017,
+      "eval_steps_per_second": 5.504,
+      "step": 1070
+    },
+    {
+      "epoch": 8.374816983894583,
+      "grad_norm": 236951.1875,
+      "learning_rate": 1.5625e-05,
+      "loss": 374.7868,
+      "step": 1080
+    },
+    {
+      "epoch": 8.374816983894583,
+      "eval_runtime": 19.6187,
+      "eval_samples_per_second": 22.02,
+      "eval_steps_per_second": 5.505,
+      "step": 1080
+    },
+    {
+      "epoch": 8.452903855539287,
+      "grad_norm": 76360.2734375,
+      "learning_rate": 1.484375e-05,
+      "loss": 312.2099,
+      "step": 1090
+    },
+    {
+      "epoch": 8.452903855539287,
+      "eval_runtime": 19.6209,
+      "eval_samples_per_second": 22.017,
+      "eval_steps_per_second": 5.504,
+      "step": 1090
+    },
+    {
+      "epoch": 8.530990727183992,
+      "grad_norm": 76876.0859375,
+      "learning_rate": 1.4062500000000001e-05,
+      "loss": 303.1329,
+      "step": 1100
+    },
+    {
+      "epoch": 8.530990727183992,
+      "eval_runtime": 19.6126,
+      "eval_samples_per_second": 22.027,
+      "eval_steps_per_second": 5.507,
+      "step": 1100
+    },
+    {
+      "epoch": 8.609077598828698,
+      "grad_norm": 80524.609375,
+      "learning_rate": 1.3281250000000001e-05,
+      "loss": 336.2521,
+      "step": 1110
+    },
+    {
+      "epoch": 8.609077598828698,
+      "eval_runtime": 19.5907,
+      "eval_samples_per_second": 22.051,
+      "eval_steps_per_second": 5.513,
+      "step": 1110
+    },
+    {
+      "epoch": 8.687164470473402,
+      "grad_norm": 114438.3828125,
+      "learning_rate": 1.25e-05,
+      "loss": 342.3281,
+      "step": 1120
+    },
+    {
+      "epoch": 8.687164470473402,
+      "eval_runtime": 19.614,
+      "eval_samples_per_second": 22.025,
+      "eval_steps_per_second": 5.506,
+      "step": 1120
+    },
+    {
+      "epoch": 8.765251342118106,
+      "grad_norm": 281197.375,
+      "learning_rate": 1.171875e-05,
+      "loss": 358.492,
+      "step": 1130
+    },
+    {
+      "epoch": 8.765251342118106,
+      "eval_runtime": 19.5986,
+      "eval_samples_per_second": 22.042,
+      "eval_steps_per_second": 5.511,
+      "step": 1130
+    },
+    {
+      "epoch": 8.843338213762811,
+      "grad_norm": 39132.8515625,
+      "learning_rate": 1.09375e-05,
+      "loss": 297.0417,
+      "step": 1140
+    },
+    {
+      "epoch": 8.843338213762811,
+      "eval_runtime": 19.6013,
+      "eval_samples_per_second": 22.039,
+      "eval_steps_per_second": 5.51,
+      "step": 1140
+    },
+    {
+      "epoch": 8.921425085407517,
+      "grad_norm": 270795.65625,
+      "learning_rate": 1.0156250000000001e-05,
+      "loss": 306.2402,
+      "step": 1150
+    },
+    {
+      "epoch": 8.921425085407517,
+      "eval_runtime": 19.6264,
+      "eval_samples_per_second": 22.011,
+      "eval_steps_per_second": 5.503,
+      "step": 1150
+    },
+    {
+      "epoch": 8.99951195705222,
+      "grad_norm": 124614.0390625,
+      "learning_rate": 9.375000000000001e-06,
+      "loss": 354.61,
+      "step": 1160
+    },
+    {
+      "epoch": 8.99951195705222,
+      "eval_runtime": 19.6258,
+      "eval_samples_per_second": 22.012,
+      "eval_steps_per_second": 5.503,
+      "step": 1160
     }
   ],
   "logging_steps": 10,