Training in progress, step 500, checkpoint

checkpoint-500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4ebfea05067cfb48f05f24a460e1d4fb4cc6d2d9843e28d35b1552b3d2945fc6
+oid sha256:631374ac29786ca2056b2753ff78b89245e52ff984e0c326635226287d64b821
 size 577789320

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c7475c3d200e582e1ef81f25d3e159868551b2d0d90c63693c9fff05860e3b16
+oid sha256:f271a371ec0c1bfd8bf4e60a22097ba7d3325af3a29954b9e83065bf445a9860
 size 1155772233

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0ad21c799b2ed3c65277e9da23ef62be7e8ff566b02f6b4d088a34a491ed6b0c
+oid sha256:59775bff66d3fdb37f6f6aec0176b9665c1c29a1d682f60e7f203aceb44c27d3
 size 14244

checkpoint-500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4a418aa2b52ddc8a5687aa016f6f0ba04d28b8859e9c0e71842a07e583777921
+oid sha256:e2c37dca05a5d090d2b76bdebe256106d41285b220290da181411a0928c0e7a2
 size 1064

checkpoint-500/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.5903987884521484,
+  "best_metric": 0.6029486060142517,
   "best_model_checkpoint": "mikhail-panzo/zlm_b64_le5_s8000/checkpoint-500",
-  "epoch": 0.418848167539267,
+  "epoch": 0.4187604690117253,
   "eval_steps": 500,
   "global_step": 500,
   "is_hyper_param_search": false,
@@ -9,81 +9,81 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.041884816753926704,
-      "grad_norm": 11.218673706054688,
-      "learning_rate": 2.4500000000000004e-07,
-      "loss": 1.1196,
+      "epoch": 0.04187604690117253,
+      "grad_norm": 5.415271759033203,
+      "learning_rate": 2.4000000000000003e-07,
+      "loss": 1.1819,
       "step": 50
     },
     {
-      "epoch": 0.08376963350785341,
-      "grad_norm": 5.423076629638672,
-      "learning_rate": 4.95e-07,
-      "loss": 1.057,
+      "epoch": 0.08375209380234507,
+      "grad_norm": 16.256921768188477,
+      "learning_rate": 4.900000000000001e-07,
+      "loss": 1.0035,
       "step": 100
     },
     {
-      "epoch": 0.1256544502617801,
-      "grad_norm": 12.713382720947266,
-      "learning_rate": 7.4e-07,
-      "loss": 1.0037,
+      "epoch": 0.12562814070351758,
+      "grad_norm": 3.2439260482788086,
+      "learning_rate": 7.350000000000001e-07,
+      "loss": 0.9742,
       "step": 150
     },
     {
-      "epoch": 0.16753926701570682,
-      "grad_norm": 4.434875965118408,
-      "learning_rate": 9.9e-07,
-      "loss": 0.9381,
+      "epoch": 0.16750418760469013,
+      "grad_norm": 4.24526834487915,
+      "learning_rate": 9.85e-07,
+      "loss": 0.9361,
       "step": 200
     },
     {
-      "epoch": 0.2094240837696335,
-      "grad_norm": 3.668553590774536,
-      "learning_rate": 1.2400000000000002e-06,
-      "loss": 0.8591,
+      "epoch": 0.20938023450586266,
+      "grad_norm": 3.625537872314453,
+      "learning_rate": 1.235e-06,
+      "loss": 0.8733,
       "step": 250
     },
     {
-      "epoch": 0.2513089005235602,
-      "grad_norm": 2.2456483840942383,
-      "learning_rate": 1.4900000000000001e-06,
-      "loss": 0.8262,
+      "epoch": 0.25125628140703515,
+      "grad_norm": 4.1455464363098145,
+      "learning_rate": 1.485e-06,
+      "loss": 0.8471,
       "step": 300
     },
     {
-      "epoch": 0.2931937172774869,
-      "grad_norm": 2.70170259475708,
-      "learning_rate": 1.74e-06,
-      "loss": 0.7634,
+      "epoch": 0.2931323283082077,
+      "grad_norm": 3.164158582687378,
+      "learning_rate": 1.7350000000000001e-06,
+      "loss": 0.838,
       "step": 350
     },
     {
-      "epoch": 0.33507853403141363,
-      "grad_norm": 2.9946064949035645,
-      "learning_rate": 1.9900000000000004e-06,
-      "loss": 0.7336,
+      "epoch": 0.33500837520938026,
+      "grad_norm": 4.24931001663208,
+      "learning_rate": 1.985e-06,
+      "loss": 0.7467,
       "step": 400
     },
     {
-      "epoch": 0.3769633507853403,
-      "grad_norm": 2.2950398921966553,
-      "learning_rate": 2.24e-06,
-      "loss": 0.7121,
+      "epoch": 0.3768844221105528,
+      "grad_norm": 2.332688331604004,
+      "learning_rate": 2.235e-06,
+      "loss": 0.7389,
       "step": 450
     },
     {
-      "epoch": 0.418848167539267,
-      "grad_norm": 8.007049560546875,
-      "learning_rate": 2.4900000000000003e-06,
-      "loss": 0.6879,
+      "epoch": 0.4187604690117253,
+      "grad_norm": 2.370961904525757,
+      "learning_rate": 2.4850000000000003e-06,
+      "loss": 0.7074,
       "step": 500
     },
     {
-      "epoch": 0.418848167539267,
-      "eval_loss": 0.5903987884521484,
-      "eval_runtime": 258.509,
-      "eval_samples_per_second": 32.838,
-      "eval_steps_per_second": 4.108,
+      "epoch": 0.4187604690117253,
+      "eval_loss": 0.6029486060142517,
+      "eval_runtime": 220.2253,
+      "eval_samples_per_second": 38.547,
+      "eval_steps_per_second": 4.822,
       "step": 500
     }
   ],
@@ -104,8 +104,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 4446942197892480.0,
-  "train_batch_size": 16,
+  "total_flos": 4787434919241216.0,
+  "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:6cbb0736de458ba42fdcabf5316a533e48ea0601626bceabd95443adeda6296a
+oid sha256:6248ade84a48a52ce5a07ff6e203bd335b96af70b5ce5ebd92ad797d93be969c
 size 5304