Training in progress, step 500, checkpoint

checkpoint-500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:418396f8389fe026d53b0494443e277c0ef031af0d5207d7f563038d96a11e8c
+oid sha256:1db3672e7cfaabe5b1f1714d79fde6df55e5e081d6358cf25c88a9833f7965f7
 size 577789320

checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7c9ba12e55f6fc8c998050abfb66fc3df21fffa86f64d4ab336163378afd7df9
+oid sha256:4014ea62140b313b8d356f4773d24ffcf6de9baaf30fb1b20355b23f4cc90326
 size 1155772233

checkpoint-500/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:cc265be82a042e4e4678eb6a1fe78828a97d6253c394371df497d3fd08de02bf
+oid sha256:a3790a7f6c80342b338faa5af1fc505f02008559f52b6506b424f6f7dd0444c0
 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": 0.5937426090240479,
-  "best_model_checkpoint": "mikhail_panzo/zlm_b64_le5_s12000/checkpoint-500",
-  "epoch": 0.418848167539267,
+  "best_metric": 0.6066474914550781,
+  "best_model_checkpoint": "mikhail-panzo/zlm_b64_le5_s12000/checkpoint-500",
+  "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": 12.821951866149902,
-      "learning_rate": 2.4000000000000003e-07,
-      "loss": 1.1152,
+      "epoch": 0.04187604690117253,
+      "grad_norm": 4.788218975067139,
+      "learning_rate": 2.4500000000000004e-07,
+      "loss": 1.1819,
       "step": 50
     },
     {
-      "epoch": 0.08376963350785341,
-      "grad_norm": 6.186221122741699,
-      "learning_rate": 4.900000000000001e-07,
-      "loss": 1.0584,
+      "epoch": 0.08375209380234507,
+      "grad_norm": 15.618935585021973,
+      "learning_rate": 4.95e-07,
+      "loss": 1.0042,
       "step": 100
     },
     {
-      "epoch": 0.1256544502617801,
-      "grad_norm": 13.08674430847168,
+      "epoch": 0.12562814070351758,
+      "grad_norm": 2.7821555137634277,
       "learning_rate": 7.4e-07,
-      "loss": 0.9996,
+      "loss": 0.9676,
       "step": 150
     },
     {
-      "epoch": 0.16753926701570682,
-      "grad_norm": 4.10632848739624,
-      "learning_rate": 9.9e-07,
-      "loss": 0.9355,
+      "epoch": 0.16750418760469013,
+      "grad_norm": 3.759263038635254,
+      "learning_rate": 9.85e-07,
+      "loss": 0.9362,
       "step": 200
     },
     {
-      "epoch": 0.2094240837696335,
-      "grad_norm": 3.2428064346313477,
-      "learning_rate": 1.2400000000000002e-06,
-      "loss": 0.852,
+      "epoch": 0.20938023450586266,
+      "grad_norm": 3.906439781188965,
+      "learning_rate": 1.235e-06,
+      "loss": 0.8667,
       "step": 250
     },
     {
-      "epoch": 0.2513089005235602,
-      "grad_norm": 2.0861401557922363,
-      "learning_rate": 1.4900000000000001e-06,
-      "loss": 0.8241,
+      "epoch": 0.25125628140703515,
+      "grad_norm": 4.397643566131592,
+      "learning_rate": 1.485e-06,
+      "loss": 0.8455,
       "step": 300
     },
     {
-      "epoch": 0.2931937172774869,
-      "grad_norm": 8.602317810058594,
-      "learning_rate": 1.74e-06,
-      "loss": 0.7629,
+      "epoch": 0.2931323283082077,
+      "grad_norm": 2.94077467918396,
+      "learning_rate": 1.7350000000000001e-06,
+      "loss": 0.8376,
       "step": 350
     },
     {
-      "epoch": 0.33507853403141363,
-      "grad_norm": 2.7594990730285645,
-      "learning_rate": 1.9900000000000004e-06,
-      "loss": 0.7418,
+      "epoch": 0.33500837520938026,
+      "grad_norm": 4.161074161529541,
+      "learning_rate": 1.985e-06,
+      "loss": 0.7521,
       "step": 400
     },
     {
-      "epoch": 0.3769633507853403,
-      "grad_norm": 2.7047348022460938,
-      "learning_rate": 2.24e-06,
-      "loss": 0.7166,
+      "epoch": 0.3768844221105528,
+      "grad_norm": 2.8088552951812744,
+      "learning_rate": 2.235e-06,
+      "loss": 0.7444,
       "step": 450
     },
     {
-      "epoch": 0.418848167539267,
-      "grad_norm": 2.0666158199310303,
-      "learning_rate": 2.4900000000000003e-06,
-      "loss": 0.6876,
+      "epoch": 0.4187604690117253,
+      "grad_norm": 2.62283992767334,
+      "learning_rate": 2.4850000000000003e-06,
+      "loss": 0.7129,
       "step": 500
     },
     {
-      "epoch": 0.418848167539267,
-      "eval_loss": 0.5937426090240479,
-      "eval_runtime": 263.0955,
-      "eval_samples_per_second": 32.266,
-      "eval_steps_per_second": 4.037,
+      "epoch": 0.4187604690117253,
+      "eval_loss": 0.6066474914550781,
+      "eval_runtime": 213.4911,
+      "eval_samples_per_second": 39.763,
+      "eval_steps_per_second": 4.974,
       "step": 500
     }
   ],
@@ -104,8 +104,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 4503134375055360.0,
-  "train_batch_size": 16,
+  "total_flos": 4878617293034496.0,
+  "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:b20f3b06aa67136878a19fd6f17a63e91db785ca54662fa7debf3d46f1fe910d
+oid sha256:f4687cdb2b9124db8f1d7d7525bb9403c3df24fb82ca7d74640c10781e1e94be
 size 5304