Training in progress, step 1000, checkpoint

checkpoint-1000/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e29ee8246b1bc22719673cb087c2b2c7035f885081e1ac1c34d17546b3821a47
+oid sha256:df681c24f5e239e07a89b1afc9241027deec1881dc507b15b9cc8bf43a206d4f
 size 577789320

checkpoint-1000/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3b0c6517b3a45b7639e34ea8f96e07188e3e8eadaeb29c06ae285da5a66de8bc
+oid sha256:c7bffed9bfc47c35a093979f162da6afcd68fe4418c55890e40dabede3b9d401
 size 1155772233

checkpoint-1000/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e8b323d971557d5fde4cc0d7edcc6e6233b739826529f518d746969aa6eb8c25
+oid sha256:cad86653bee64e341a6febb0b6b2c0af9e7a5388df880cf238387c12cfa6a910
 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": 0.4933945834636688,
-  "best_model_checkpoint": "mikhail_panzo/zlm_b64_le5_s12000/checkpoint-1000",
-  "epoch": 0.837696335078534,
+  "best_metric": 0.4979843497276306,
+  "best_model_checkpoint": "mikhail-panzo/zlm_b64_le5_s12000/checkpoint-1000",
+  "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": 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
     },
     {
-      "epoch": 0.4607329842931937,
-      "grad_norm": 2.955193519592285,
-      "learning_rate": 2.7400000000000004e-06,
-      "loss": 0.673,
+      "epoch": 0.46063651591289784,
+      "grad_norm": 3.654665946960449,
+      "learning_rate": 2.7350000000000005e-06,
+      "loss": 0.6772,
       "step": 550
     },
     {
-      "epoch": 0.5026178010471204,
-      "grad_norm": 4.856322765350342,
-      "learning_rate": 2.99e-06,
-      "loss": 0.6539,
+      "epoch": 0.5025125628140703,
+      "grad_norm": 5.1851887702941895,
+      "learning_rate": 2.9850000000000002e-06,
+      "loss": 0.6552,
       "step": 600
     },
     {
-      "epoch": 0.5445026178010471,
-      "grad_norm": 7.631103992462158,
-      "learning_rate": 3.2400000000000003e-06,
-      "loss": 0.6588,
+      "epoch": 0.5443886097152428,
+      "grad_norm": 4.017187595367432,
+      "learning_rate": 3.2350000000000004e-06,
+      "loss": 0.6506,
       "step": 650
     },
     {
-      "epoch": 0.5863874345549738,
-      "grad_norm": 2.715771436691284,
-      "learning_rate": 3.49e-06,
-      "loss": 0.624,
+      "epoch": 0.5862646566164154,
+      "grad_norm": 2.1161563396453857,
+      "learning_rate": 3.485e-06,
+      "loss": 0.6301,
       "step": 700
     },
     {
-      "epoch": 0.6282722513089005,
-      "grad_norm": 2.7684273719787598,
-      "learning_rate": 3.74e-06,
-      "loss": 0.6116,
+      "epoch": 0.628140703517588,
+      "grad_norm": 1.857313871383667,
+      "learning_rate": 3.7350000000000002e-06,
+      "loss": 0.6373,
       "step": 750
     },
     {
-      "epoch": 0.6701570680628273,
-      "grad_norm": 2.219599723815918,
-      "learning_rate": 3.990000000000001e-06,
-      "loss": 0.6097,
+      "epoch": 0.6700167504187605,
+      "grad_norm": 3.6841330528259277,
+      "learning_rate": 3.985000000000001e-06,
+      "loss": 0.6305,
       "step": 800
     },
     {
-      "epoch": 0.7120418848167539,
-      "grad_norm": 3.749188184738159,
-      "learning_rate": 4.24e-06,
-      "loss": 0.614,
+      "epoch": 0.711892797319933,
+      "grad_norm": 2.6450912952423096,
+      "learning_rate": 4.235e-06,
+      "loss": 0.6169,
       "step": 850
     },
     {
-      "epoch": 0.7539267015706806,
-      "grad_norm": 2.120619773864746,
-      "learning_rate": 4.49e-06,
-      "loss": 0.5814,
+      "epoch": 0.7537688442211056,
+      "grad_norm": 1.7241184711456299,
+      "learning_rate": 4.485e-06,
+      "loss": 0.5933,
       "step": 900
     },
     {
-      "epoch": 0.7958115183246073,
-      "grad_norm": 2.7725327014923096,
-      "learning_rate": 4.74e-06,
-      "loss": 0.5895,
+      "epoch": 0.7956448911222781,
+      "grad_norm": 2.602735996246338,
+      "learning_rate": 4.735e-06,
+      "loss": 0.583,
       "step": 950
     },
     {
-      "epoch": 0.837696335078534,
-      "grad_norm": 9.851961135864258,
-      "learning_rate": 4.9900000000000005e-06,
-      "loss": 0.5623,
+      "epoch": 0.8375209380234506,
+      "grad_norm": 12.419028282165527,
+      "learning_rate": 4.9850000000000006e-06,
+      "loss": 0.5922,
       "step": 1000
     },
     {
-      "epoch": 0.837696335078534,
-      "eval_loss": 0.4933945834636688,
-      "eval_runtime": 258.739,
-      "eval_samples_per_second": 32.809,
-      "eval_steps_per_second": 4.105,
+      "epoch": 0.8375209380234506,
+      "eval_loss": 0.4979843497276306,
+      "eval_runtime": 214.0502,
+      "eval_samples_per_second": 39.659,
+      "eval_steps_per_second": 4.961,
       "step": 1000
     }
   ],
@@ -182,8 +182,8 @@
       "attributes": {}
     }
   },
-  "total_flos": 9000282308548608.0,
-  "train_batch_size": 16,
+  "total_flos": 9749197233971712.0,
+  "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:b20f3b06aa67136878a19fd6f17a63e91db785ca54662fa7debf3d46f1fe910d
+oid sha256:f4687cdb2b9124db8f1d7d7525bb9403c3df24fb82ca7d74640c10781e1e94be
 size 5304