Training in progress, step 1000, checkpoint

checkpoint-1000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:537a22a69200ee35e29b60f5692289461a6733be3a1534af4430c82a3f9f23a8
+oid sha256:90d986cf1255b27ffc3350ae3f307b4f1ca147e2f5fa37b51b5ddcc04d55e548
 size 577789320

checkpoint-1000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a497454262ba6e4ca9a8c82617ae1fe128f1e5384b0ebecdce2472827b197b28
+oid sha256:0dbd2f9bda20d2c03463ef8f2cb6928197337f4b6571ecae87ad4a0494eeca38
 size 1155772233

checkpoint-1000/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:db5636898e47dba496c436a5e94385aeb16d7310f5a73718689dab9ea5000167
+oid sha256:ca7d5ee2051d64ea6122bf3bb55105a3295eac4cd2ed5d3ee59c6391d438e17b
 size 14244

checkpoint-1000/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2599834a5bb25566393aaf9a393926be754987442fa6dd0ec6e936958ebe0844
+oid sha256:d01ebd56c5cf26ace306036e45603a144e0f77296a474e251c4c195ea1480566
 size 1064

checkpoint-1000/trainer_state.json

@@ -1,7 +1,7 @@
 {
-  "best_metric": 0.49923568964004517,
+  "best_metric": 0.49684029817581177,
   "best_model_checkpoint": "mikhail-panzo/zlm_b64_le5_s8000/checkpoint-1000",
-  "epoch": 0.837696335078534,
+  "epoch": 0.8375209380234506,
   "eval_steps": 500,
   "global_step": 1000,
   "is_hyper_param_search": false,
@@ -9,159 +9,159 @@
   "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
     },
     {
-      "epoch": 0.4607329842931937,
-      "grad_norm": 3.0767672061920166,
-      "learning_rate": 2.7400000000000004e-06,
-      "loss": 0.6734,
+      "epoch": 0.46063651591289784,
+      "grad_norm": 14.466177940368652,
+      "learning_rate": 2.7350000000000005e-06,
+      "loss": 0.6738,
       "step": 550
     },
     {
-      "epoch": 0.5026178010471204,
-      "grad_norm": 8.150467872619629,
-      "learning_rate": 2.99e-06,
-      "loss": 0.6539,
+      "epoch": 0.5025125628140703,
+      "grad_norm": 3.282449722290039,
+      "learning_rate": 2.9850000000000002e-06,
+      "loss": 0.6517,
       "step": 600
     },
     {
-      "epoch": 0.5445026178010471,
-      "grad_norm": 2.2473220825195312,
-      "learning_rate": 3.2400000000000003e-06,
-      "loss": 0.657,
+      "epoch": 0.5443886097152428,
+      "grad_norm": 2.621645212173462,
+      "learning_rate": 3.2350000000000004e-06,
+      "loss": 0.6479,
       "step": 650
     },
     {
-      "epoch": 0.5863874345549738,
-      "grad_norm": 2.448143482208252,
-      "learning_rate": 3.49e-06,
-      "loss": 0.6258,
+      "epoch": 0.5862646566164154,
+      "grad_norm": 2.4493019580841064,
+      "learning_rate": 3.485e-06,
+      "loss": 0.6321,
       "step": 700
     },
     {
-      "epoch": 0.6282722513089005,
-      "grad_norm": 4.810636520385742,
-      "learning_rate": 3.74e-06,
-      "loss": 0.6155,
+      "epoch": 0.628140703517588,
+      "grad_norm": 1.919048547744751,
+      "learning_rate": 3.7350000000000002e-06,
+      "loss": 0.6407,
       "step": 750
     },
     {
-      "epoch": 0.6701570680628273,
-      "grad_norm": 2.5927581787109375,
-      "learning_rate": 3.990000000000001e-06,
-      "loss": 0.6073,
+      "epoch": 0.6700167504187605,
+      "grad_norm": 2.8260602951049805,
+      "learning_rate": 3.985000000000001e-06,
+      "loss": 0.6279,
       "step": 800
     },
     {
-      "epoch": 0.7120418848167539,
-      "grad_norm": 1.913765549659729,
-      "learning_rate": 4.24e-06,
-      "loss": 0.6118,
+      "epoch": 0.711892797319933,
+      "grad_norm": 3.333841323852539,
+      "learning_rate": 4.235e-06,
+      "loss": 0.6143,
       "step": 850
     },
     {
-      "epoch": 0.7539267015706806,
-      "grad_norm": 2.210376262664795,
-      "learning_rate": 4.49e-06,
-      "loss": 0.5807,
+      "epoch": 0.7537688442211056,
+      "grad_norm": 1.7883094549179077,
+      "learning_rate": 4.485e-06,
+      "loss": 0.5949,
       "step": 900
     },
     {
-      "epoch": 0.7958115183246073,
-      "grad_norm": 3.323096752166748,
-      "learning_rate": 4.74e-06,
-      "loss": 0.5915,
+      "epoch": 0.7956448911222781,
+      "grad_norm": 2.48634934425354,
+      "learning_rate": 4.735e-06,
+      "loss": 0.5875,
       "step": 950
     },
     {
-      "epoch": 0.837696335078534,
-      "grad_norm": 2.1455421447753906,
-      "learning_rate": 4.9900000000000005e-06,
-      "loss": 0.5606,
+      "epoch": 0.8375209380234506,
+      "grad_norm": 5.4277873039245605,
+      "learning_rate": 4.9850000000000006e-06,
+      "loss": 0.5916,
       "step": 1000
     },
     {
-      "epoch": 0.837696335078534,
-      "eval_loss": 0.49923568964004517,
-      "eval_runtime": 260.1008,
-      "eval_samples_per_second": 32.637,
-      "eval_steps_per_second": 4.083,
+      "epoch": 0.8375209380234506,
+      "eval_loss": 0.49684029817581177,
+      "eval_runtime": 219.2621,
+      "eval_samples_per_second": 38.716,
+      "eval_steps_per_second": 4.844,
       "step": 1000
     }
   ],
@@ -182,8 +182,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 8963407491426432.0,
-  "train_batch_size": 16,
+  "total_flos": 9705906392675328.0,
+  "train_batch_size": 32,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-1000/training_args.bin

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