Training in progress, step 500

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e14d49c984a2dd95c6f65057ebb9749f23d19ce96b7e07f88a45470915eddb75
+oid sha256:746732a955ce5ac447aae4ae74e4f595331f971b2b8b78db1b28ff7cc415b7b9
 size 268290900

run-11/checkpoint-2500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:53b16b25375514918da67404b3403c850c15a69162ae13812852117b14e664d6
+oid sha256:0c3d81cf1e94136cf69f7afe09173f6fff54ee913216fd1109c21e652a8bd239
 size 5176

run-11/checkpoint-3000/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:53b16b25375514918da67404b3403c850c15a69162ae13812852117b14e664d6
+oid sha256:0c3d81cf1e94136cf69f7afe09173f6fff54ee913216fd1109c21e652a8bd239
 size 5176

run-12/checkpoint-1000/trainer_state.json

@@ -10,50 +10,50 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.6064516129032258,
-      "eval_loss": 0.2185761034488678,
-      "eval_runtime": 1.3638,
-      "eval_samples_per_second": 2273.102,
-      "eval_steps_per_second": 47.662,
+      "eval_accuracy": 0.6732258064516129,
+      "eval_loss": 0.40791815519332886,
+      "eval_runtime": 1.3596,
+      "eval_samples_per_second": 2280.098,
+      "eval_steps_per_second": 47.809,
       "step": 318
     },
     {
       "epoch": 1.5723270440251573,
-      "grad_norm": 0.5516161918640137,
-      "learning_rate": 1.685534591194969e-05,
-      "loss": 0.3519,
+      "grad_norm": 1.0064228773117065,
+      "learning_rate": 1.650593990216632e-05,
+      "loss": 0.642,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8412903225806452,
-      "eval_loss": 0.10163559764623642,
-      "eval_runtime": 1.3673,
-      "eval_samples_per_second": 2267.277,
-      "eval_steps_per_second": 47.54,
+      "eval_accuracy": 0.8425806451612903,
+      "eval_loss": 0.14076808094978333,
+      "eval_runtime": 1.3551,
+      "eval_samples_per_second": 2287.635,
+      "eval_steps_per_second": 47.967,
       "step": 636
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8858064516129033,
-      "eval_loss": 0.06496834009885788,
-      "eval_runtime": 1.3574,
-      "eval_samples_per_second": 2283.817,
-      "eval_steps_per_second": 47.886,
+      "eval_accuracy": 0.8961290322580645,
+      "eval_loss": 0.07251902669668198,
+      "eval_runtime": 1.3507,
+      "eval_samples_per_second": 2295.17,
+      "eval_steps_per_second": 48.125,
       "step": 954
     },
     {
       "epoch": 3.1446540880503147,
-      "grad_norm": 0.5089602470397949,
-      "learning_rate": 1.371069182389937e-05,
-      "loss": 0.1182,
+      "grad_norm": 0.7264623045921326,
+      "learning_rate": 1.3011879804332637e-05,
+      "loss": 0.1644,
       "step": 1000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 3180,
+  "max_steps": 2862,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 9,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -71,8 +71,8 @@
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.86409290852114,
-    "num_train_epochs": 10,
-    "temperature": 6
+    "alpha": 0.9838674034859954,
+    "num_train_epochs": 9,
+    "temperature": 2
   }
 }

run-12/checkpoint-1500/trainer_state.json

@@ -10,66 +10,66 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.6064516129032258,
-      "eval_loss": 0.2185761034488678,
-      "eval_runtime": 1.3638,
-      "eval_samples_per_second": 2273.102,
-      "eval_steps_per_second": 47.662,
+      "eval_accuracy": 0.6732258064516129,
+      "eval_loss": 0.40791815519332886,
+      "eval_runtime": 1.3596,
+      "eval_samples_per_second": 2280.098,
+      "eval_steps_per_second": 47.809,
       "step": 318
     },
     {
       "epoch": 1.5723270440251573,
-      "grad_norm": 0.5516161918640137,
-      "learning_rate": 1.685534591194969e-05,
-      "loss": 0.3519,
+      "grad_norm": 1.0064228773117065,
+      "learning_rate": 1.650593990216632e-05,
+      "loss": 0.642,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8412903225806452,
-      "eval_loss": 0.10163559764623642,
-      "eval_runtime": 1.3673,
-      "eval_samples_per_second": 2267.277,
-      "eval_steps_per_second": 47.54,
+      "eval_accuracy": 0.8425806451612903,
+      "eval_loss": 0.14076808094978333,
+      "eval_runtime": 1.3551,
+      "eval_samples_per_second": 2287.635,
+      "eval_steps_per_second": 47.967,
       "step": 636
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8858064516129033,
-      "eval_loss": 0.06496834009885788,
-      "eval_runtime": 1.3574,
-      "eval_samples_per_second": 2283.817,
-      "eval_steps_per_second": 47.886,
+      "eval_accuracy": 0.8961290322580645,
+      "eval_loss": 0.07251902669668198,
+      "eval_runtime": 1.3507,
+      "eval_samples_per_second": 2295.17,
+      "eval_steps_per_second": 48.125,
       "step": 954
     },
     {
       "epoch": 3.1446540880503147,
-      "grad_norm": 0.5089602470397949,
-      "learning_rate": 1.371069182389937e-05,
-      "loss": 0.1182,
+      "grad_norm": 0.7264623045921326,
+      "learning_rate": 1.3011879804332637e-05,
+      "loss": 0.1644,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9003225806451612,
-      "eval_loss": 0.048449017107486725,
-      "eval_runtime": 1.3674,
-      "eval_samples_per_second": 2267.023,
-      "eval_steps_per_second": 47.534,
+      "eval_accuracy": 0.9141935483870968,
+      "eval_loss": 0.05225619301199913,
+      "eval_runtime": 1.3588,
+      "eval_samples_per_second": 2281.449,
+      "eval_steps_per_second": 47.837,
       "step": 1272
     },
     {
       "epoch": 4.716981132075472,
-      "grad_norm": 0.3282025456428528,
-      "learning_rate": 1.0566037735849058e-05,
-      "loss": 0.0719,
+      "grad_norm": 0.4414461851119995,
+      "learning_rate": 9.517819706498952e-06,
+      "loss": 0.0818,
       "step": 1500
     }
   ],
   "logging_steps": 500,
-  "max_steps": 3180,
+  "max_steps": 2862,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 9,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -87,8 +87,8 @@
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.86409290852114,
-    "num_train_epochs": 10,
-    "temperature": 6
+    "alpha": 0.9838674034859954,
+    "num_train_epochs": 9,
+    "temperature": 2
   }
 }

run-12/checkpoint-2000/trainer_state.json

@@ -10,91 +10,91 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.6064516129032258,
-      "eval_loss": 0.2185761034488678,
-      "eval_runtime": 1.3638,
-      "eval_samples_per_second": 2273.102,
-      "eval_steps_per_second": 47.662,
+      "eval_accuracy": 0.6732258064516129,
+      "eval_loss": 0.40791815519332886,
+      "eval_runtime": 1.3596,
+      "eval_samples_per_second": 2280.098,
+      "eval_steps_per_second": 47.809,
       "step": 318
     },
     {
       "epoch": 1.5723270440251573,
-      "grad_norm": 0.5516161918640137,
-      "learning_rate": 1.685534591194969e-05,
-      "loss": 0.3519,
+      "grad_norm": 1.0064228773117065,
+      "learning_rate": 1.650593990216632e-05,
+      "loss": 0.642,
       "step": 500
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.8412903225806452,
-      "eval_loss": 0.10163559764623642,
-      "eval_runtime": 1.3673,
-      "eval_samples_per_second": 2267.277,
-      "eval_steps_per_second": 47.54,
+      "eval_accuracy": 0.8425806451612903,
+      "eval_loss": 0.14076808094978333,
+      "eval_runtime": 1.3551,
+      "eval_samples_per_second": 2287.635,
+      "eval_steps_per_second": 47.967,
       "step": 636
     },
     {
       "epoch": 3.0,
-      "eval_accuracy": 0.8858064516129033,
-      "eval_loss": 0.06496834009885788,
-      "eval_runtime": 1.3574,
-      "eval_samples_per_second": 2283.817,
-      "eval_steps_per_second": 47.886,
+      "eval_accuracy": 0.8961290322580645,
+      "eval_loss": 0.07251902669668198,
+      "eval_runtime": 1.3507,
+      "eval_samples_per_second": 2295.17,
+      "eval_steps_per_second": 48.125,
       "step": 954
     },
     {
       "epoch": 3.1446540880503147,
-      "grad_norm": 0.5089602470397949,
-      "learning_rate": 1.371069182389937e-05,
-      "loss": 0.1182,
+      "grad_norm": 0.7264623045921326,
+      "learning_rate": 1.3011879804332637e-05,
+      "loss": 0.1644,
       "step": 1000
     },
     {
       "epoch": 4.0,
-      "eval_accuracy": 0.9003225806451612,
-      "eval_loss": 0.048449017107486725,
-      "eval_runtime": 1.3674,
-      "eval_samples_per_second": 2267.023,
-      "eval_steps_per_second": 47.534,
+      "eval_accuracy": 0.9141935483870968,
+      "eval_loss": 0.05225619301199913,
+      "eval_runtime": 1.3588,
+      "eval_samples_per_second": 2281.449,
+      "eval_steps_per_second": 47.837,
       "step": 1272
     },
     {
       "epoch": 4.716981132075472,
-      "grad_norm": 0.3282025456428528,
-      "learning_rate": 1.0566037735849058e-05,
-      "loss": 0.0719,
+      "grad_norm": 0.4414461851119995,
+      "learning_rate": 9.517819706498952e-06,
+      "loss": 0.0818,
       "step": 1500
     },
     {
       "epoch": 5.0,
-      "eval_accuracy": 0.9135483870967742,
-      "eval_loss": 0.03883416950702667,
-      "eval_runtime": 1.3721,
-      "eval_samples_per_second": 2259.288,
-      "eval_steps_per_second": 47.372,
+      "eval_accuracy": 0.9238709677419354,
+      "eval_loss": 0.04304172471165657,
+      "eval_runtime": 1.3623,
+      "eval_samples_per_second": 2275.546,
+      "eval_steps_per_second": 47.713,
       "step": 1590
     },
     {
       "epoch": 6.0,
-      "eval_accuracy": 0.922258064516129,
-      "eval_loss": 0.03377517685294151,
-      "eval_runtime": 1.5655,
-      "eval_samples_per_second": 1980.228,
-      "eval_steps_per_second": 41.521,
+      "eval_accuracy": 0.9283870967741935,
+      "eval_loss": 0.038618117570877075,
+      "eval_runtime": 1.3565,
+      "eval_samples_per_second": 2285.293,
+      "eval_steps_per_second": 47.917,
       "step": 1908
     },
     {
       "epoch": 6.289308176100629,
-      "grad_norm": 0.24476511776447296,
-      "learning_rate": 7.421383647798742e-06,
-      "loss": 0.0554,
+      "grad_norm": 0.3635008931159973,
+      "learning_rate": 6.02375960866527e-06,
+      "loss": 0.0619,
       "step": 2000
     }
   ],
   "logging_steps": 500,
-  "max_steps": 3180,
+  "max_steps": 2862,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 9,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -112,8 +112,8 @@
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.86409290852114,
-    "num_train_epochs": 10,
-    "temperature": 6
+    "alpha": 0.9838674034859954,
+    "num_train_epochs": 9,
+    "temperature": 2
   }
 }

run-12/checkpoint-500/model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:52acb0f10dbc4d7d9fc1bdc8234f938bf09a424cb2ae024de0d4b496ce371435
+oid sha256:746732a955ce5ac447aae4ae74e4f595331f971b2b8b78db1b28ff7cc415b7b9
 size 268290900

run-12/checkpoint-500/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ede2108aa21755bea146deda4c9c419892c1cd9d89a6a7f459b3be1057346deb
+oid sha256:9705b8a6db729e9a5e87252472c23151c8c318a4f1c6679911400c9b1a41b495
 size 536643898

run-12/checkpoint-500/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:04366f62f8f88f5a8265df59adb051b320463277845db80e7fa43f13110c18c9
+oid sha256:98041bd7cae455426e290a1a0ee683bd5dd30893f7451fec3a464ae8995b17e4
 size 1064

run-12/checkpoint-500/trainer_state.json

@@ -10,25 +10,25 @@
   "log_history": [
     {
       "epoch": 1.0,
-      "eval_accuracy": 0.6064516129032258,
-      "eval_loss": 0.2185761034488678,
-      "eval_runtime": 1.3638,
-      "eval_samples_per_second": 2273.102,
-      "eval_steps_per_second": 47.662,
+      "eval_accuracy": 0.6732258064516129,
+      "eval_loss": 0.40791815519332886,
+      "eval_runtime": 1.3596,
+      "eval_samples_per_second": 2280.098,
+      "eval_steps_per_second": 47.809,
       "step": 318
     },
     {
       "epoch": 1.5723270440251573,
-      "grad_norm": 0.5516161918640137,
-      "learning_rate": 1.685534591194969e-05,
-      "loss": 0.3519,
+      "grad_norm": 1.0064228773117065,
+      "learning_rate": 1.650593990216632e-05,
+      "loss": 0.642,
       "step": 500
     }
   ],
   "logging_steps": 500,
-  "max_steps": 3180,
+  "max_steps": 2862,
   "num_input_tokens_seen": 0,
-  "num_train_epochs": 10,
+  "num_train_epochs": 9,
   "save_steps": 500,
   "stateful_callbacks": {
     "TrainerControl": {
@@ -46,8 +46,8 @@
   "train_batch_size": 48,
   "trial_name": null,
   "trial_params": {
-    "alpha": 0.86409290852114,
-    "num_train_epochs": 10,
-    "temperature": 6
+    "alpha": 0.9838674034859954,
+    "num_train_epochs": 9,
+    "temperature": 2
   }
 }

run-12/checkpoint-500/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec0cc6acf298edb76f3b5ee73f8edb2020abda4adeb81e0f2c668de6a46acc2f
+oid sha256:825ce3dae72879e90f1e7c8bf9db5978e39cf9b4e9cf4d01acbdb15f1075ac03
 size 5176

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0c3d81cf1e94136cf69f7afe09173f6fff54ee913216fd1109c21e652a8bd239
+oid sha256:825ce3dae72879e90f1e7c8bf9db5978e39cf9b4e9cf4d01acbdb15f1075ac03
 size 5176