Upload folder using huggingface_hub

.gitattributes

@@ -53,3 +53,4 @@ checkpoints-v5.10/checkpoint-7168/eval_state.json filter=lfs diff=lfs merge=lfs
 checkpoints-v4.5/checkpoint-21504/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v5.11/checkpoint-10240/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v4.6/checkpoint-13312/eval_state.json filter=lfs diff=lfs merge=lfs -text
+checkpoints-v5.11-b/checkpoint-9216/eval_state.json filter=lfs diff=lfs merge=lfs -text

checkpoints-v5.11-b/checkpoint-9216/ema.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:04c2769d151b579dfa78b54963eb7429f42acda7ba5a5e7065038c7488c481b1
+size 57867144

checkpoints-v5.11-b/checkpoint-9216/eval_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1d4abc7fcec831f16d1211a677a4949a03a528b97a3fd93bf05927da7429ccd0
+size 60057280

checkpoints-v5.11-b/checkpoint-9216/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cab7719a2eb5ac904250ec8c346cf098ba43fa46499beae219404110d813633f
+size 57867176

checkpoints-v5.11-b/checkpoint-9216/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7316b7f6fd55c40b66f463390042e6fb8476ddfbbe0ac0fc6d639baf4049e05d
+size 83157195

checkpoints-v5.11-b/checkpoint-9216/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bb8ec39a6ee8bc717d804213ed1dac1a8c0431955a05c8fed2d232978e8d6a6b
+size 14645

checkpoints-v5.11-b/checkpoint-9216/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1c997a9c1d259a308b04d914b2cca91bc30def141028501777c50fb89f7d8a80
+size 1383

checkpoints-v5.11-b/checkpoint-9216/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e39796bb2c17e5aae66cfe6c084e5a6bc820f97a71a9e7ef4cd1e799661eb7f7
+size 1465

checkpoints-v5.11-b/checkpoint-9216/trainer_state.json

@@ -0,0 +1,376 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3971814596935807,
+  "eval_steps": 1024,
+  "global_step": 9216,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04413127329928675,
+      "grad_norm": 2.52653431892395,
+      "learning_rate": 9.990234375e-05,
+      "loss": 12.195906639099121,
+      "step": 1024
+    },
+    {
+      "epoch": 0.04413127329928675,
+      "eval_bleu": 0.20492372710719675,
+      "eval_ce_clean_loss": 1.7283075054062964,
+      "eval_ce_pred_loss": 4.658400398327598,
+      "eval_flow_consistency_loss": 0.1969631168443257,
+      "eval_flow_mse_loss": 0.8933940772562902,
+      "eval_loss": 5.981063390337328,
+      "flow/cos_sim": 0.40864291520261053,
+      "flow/improvement_ratio": 0.974421706789338,
+      "flow/mag_ratio_mean": 0.3482704608679326,
+      "flow/mag_ratio_std": 0.11673642702893153,
+      "step": 1024
+    },
+    {
+      "epoch": 0.04413127329928675,
+      "eval_bleu": 0.20492372710719675,
+      "eval_ce_clean_loss": 1.7283075054062964,
+      "eval_ce_pred_loss": 4.658400398327598,
+      "eval_flow_consistency_loss": 0.1969631168443257,
+      "eval_flow_mse_loss": 0.8933940772562902,
+      "eval_loss": 5.981063390337328,
+      "eval_runtime": 220.5442,
+      "eval_samples_per_second": 136.027,
+      "eval_steps_per_second": 2.127,
+      "flow/cos_sim": 0.40864291520261053,
+      "flow/improvement_ratio": 0.974421706789338,
+      "flow/mag_ratio_mean": 0.3482704608679326,
+      "flow/mag_ratio_std": 0.11673642702893153,
+      "step": 1024
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "grad_norm": 0.8477205038070679,
+      "learning_rate": 9.9476028157316e-05,
+      "loss": 4.296904563903809,
+      "step": 2048
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "eval_bleu": 0.4203954756104412,
+      "eval_ce_clean_loss": 0.31376014100208977,
+      "eval_ce_pred_loss": 2.8760192099410586,
+      "eval_flow_consistency_loss": 0.2066244624698086,
+      "eval_flow_mse_loss": 0.9511578829049556,
+      "eval_loss": 3.3814436672592976,
+      "flow/cos_sim": 0.5975300241380866,
+      "flow/improvement_ratio": 0.9937155933014111,
+      "flow/mag_ratio_mean": 0.5349513870566639,
+      "flow/mag_ratio_std": 0.10326376852831606,
+      "step": 2048
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "eval_bleu": 0.4203954756104412,
+      "eval_ce_clean_loss": 0.31376014100208977,
+      "eval_ce_pred_loss": 2.8760192099410586,
+      "eval_flow_consistency_loss": 0.2066244624698086,
+      "eval_flow_mse_loss": 0.9511578829049556,
+      "eval_loss": 3.3814436672592976,
+      "eval_runtime": 220.3483,
+      "eval_samples_per_second": 136.148,
+      "eval_steps_per_second": 2.128,
+      "flow/cos_sim": 0.5975300241380866,
+      "flow/improvement_ratio": 0.9937155933014111,
+      "flow/mag_ratio_mean": 0.5349513870566639,
+      "flow/mag_ratio_std": 0.10326376852831606,
+      "step": 2048
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "grad_norm": 0.5322298407554626,
+      "learning_rate": 9.791307026072513e-05,
+      "loss": 3.1445906162261963,
+      "step": 3072
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "eval_bleu": 0.48539258855261896,
+      "eval_ce_clean_loss": 0.13052247220010899,
+      "eval_ce_pred_loss": 2.364091571205969,
+      "eval_flow_consistency_loss": 0.22353638632338185,
+      "eval_flow_mse_loss": 0.9629143555281259,
+      "eval_loss": 2.860069102569938,
+      "flow/cos_sim": 0.6811817029137601,
+      "flow/improvement_ratio": 0.9940002780479155,
+      "flow/mag_ratio_mean": 0.6364521633333234,
+      "flow/mag_ratio_std": 0.09635336341252967,
+      "step": 3072
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "eval_bleu": 0.48539258855261896,
+      "eval_ce_clean_loss": 0.13052247220010899,
+      "eval_ce_pred_loss": 2.364091571205969,
+      "eval_flow_consistency_loss": 0.22353638632338185,
+      "eval_flow_mse_loss": 0.9629143555281259,
+      "eval_loss": 2.860069102569938,
+      "eval_runtime": 219.3797,
+      "eval_samples_per_second": 136.749,
+      "eval_steps_per_second": 2.138,
+      "flow/cos_sim": 0.6811817029137601,
+      "flow/improvement_ratio": 0.9940002780479155,
+      "flow/mag_ratio_mean": 0.6364521633333234,
+      "flow/mag_ratio_std": 0.09635336341252967,
+      "step": 3072
+    },
+    {
+      "epoch": 0.176525093197147,
+      "grad_norm": 0.4867124855518341,
+      "learning_rate": 9.53439476074686e-05,
+      "loss": 2.76550555229187,
+      "step": 4096
+    },
+    {
+      "epoch": 0.176525093197147,
+      "eval_bleu": 0.5224707415420085,
+      "eval_ce_clean_loss": 0.07004804110952786,
+      "eval_ce_pred_loss": 2.1067297717909823,
+      "eval_flow_consistency_loss": 0.22862433567484305,
+      "eval_flow_mse_loss": 0.9485902956553868,
+      "eval_loss": 2.6076613288444244,
+      "flow/cos_sim": 0.7283694244651143,
+      "flow/improvement_ratio": 0.9943000164621675,
+      "flow/mag_ratio_mean": 0.693297458991313,
+      "flow/mag_ratio_std": 0.10988020541062996,
+      "step": 4096
+    },
+    {
+      "epoch": 0.176525093197147,
+      "eval_bleu": 0.5224707415420085,
+      "eval_ce_clean_loss": 0.07004804110952786,
+      "eval_ce_pred_loss": 2.1067297717909823,
+      "eval_flow_consistency_loss": 0.22862433567484305,
+      "eval_flow_mse_loss": 0.9485902956553868,
+      "eval_loss": 2.6076613288444244,
+      "eval_runtime": 220.6021,
+      "eval_samples_per_second": 135.991,
+      "eval_steps_per_second": 2.126,
+      "flow/cos_sim": 0.7283694244651143,
+      "flow/improvement_ratio": 0.9943000164621675,
+      "flow/mag_ratio_mean": 0.693297458991313,
+      "flow/mag_ratio_std": 0.10988020541062996,
+      "step": 4096
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "grad_norm": 1.0894174575805664,
+      "learning_rate": 9.182261125213742e-05,
+      "loss": 2.59098219871521,
+      "step": 5120
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "eval_bleu": 0.5393664611968214,
+      "eval_ce_clean_loss": 0.042873796897688145,
+      "eval_ce_pred_loss": 1.9514871731495806,
+      "eval_flow_consistency_loss": 0.21636765890284135,
+      "eval_flow_mse_loss": 0.9199989291904832,
+      "eval_loss": 2.43709755960558,
+      "flow/cos_sim": 0.7597128864544541,
+      "flow/improvement_ratio": 0.9947783349673631,
+      "flow/mag_ratio_mean": 0.7298253399731,
+      "flow/mag_ratio_std": 0.11818532291442346,
+      "step": 5120
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "eval_bleu": 0.5393664611968214,
+      "eval_ce_clean_loss": 0.042873796897688145,
+      "eval_ce_pred_loss": 1.9514871731495806,
+      "eval_flow_consistency_loss": 0.21636765890284135,
+      "eval_flow_mse_loss": 0.9199989291904832,
+      "eval_loss": 2.43709755960558,
+      "eval_runtime": 221.6485,
+      "eval_samples_per_second": 135.349,
+      "eval_steps_per_second": 2.116,
+      "flow/cos_sim": 0.7597128864544541,
+      "flow/improvement_ratio": 0.9947783349673631,
+      "flow/mag_ratio_mean": 0.7298253399731,
+      "flow/mag_ratio_std": 0.11818532291442346,
+      "step": 5120
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "grad_norm": 0.6129837036132812,
+      "learning_rate": 8.742300854391668e-05,
+      "loss": 2.4653327465057373,
+      "step": 6144
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "eval_bleu": 0.5560727460810808,
+      "eval_ce_clean_loss": 0.028736061462238907,
+      "eval_ce_pred_loss": 1.8302468360105812,
+      "eval_flow_consistency_loss": 0.19070348343742427,
+      "eval_flow_mse_loss": 0.9170380932435807,
+      "eval_loss": 2.3222986602071507,
+      "flow/cos_sim": 0.785038922641323,
+      "flow/improvement_ratio": 0.9949500379023521,
+      "flow/mag_ratio_mean": 0.7534581313509423,
+      "flow/mag_ratio_std": 0.1224009687267641,
+      "step": 6144
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "eval_bleu": 0.5560727460810808,
+      "eval_ce_clean_loss": 0.028736061462238907,
+      "eval_ce_pred_loss": 1.8302468360105812,
+      "eval_flow_consistency_loss": 0.19070348343742427,
+      "eval_flow_mse_loss": 0.9170380932435807,
+      "eval_loss": 2.3222986602071507,
+      "eval_runtime": 223.1234,
+      "eval_samples_per_second": 134.455,
+      "eval_steps_per_second": 2.102,
+      "flow/cos_sim": 0.785038922641323,
+      "flow/improvement_ratio": 0.9949500379023521,
+      "flow/mag_ratio_mean": 0.7534581313509423,
+      "flow/mag_ratio_std": 0.1224009687267641,
+      "step": 6144
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "grad_norm": 0.62188321352005,
+      "learning_rate": 8.223753024725232e-05,
+      "loss": 2.393148899078369,
+      "step": 7168
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "eval_bleu": 0.564904310555124,
+      "eval_ce_clean_loss": 0.020537127653307626,
+      "eval_ce_pred_loss": 1.7842765613151257,
+      "eval_flow_consistency_loss": 0.17003079746831964,
+      "eval_flow_mse_loss": 0.9064064752826813,
+      "eval_loss": 2.2609525786788223,
+      "flow/cos_sim": 0.7994557322977957,
+      "flow/improvement_ratio": 0.9958853139551972,
+      "flow/mag_ratio_mean": 0.7681991188510903,
+      "flow/mag_ratio_std": 0.12335055292860023,
+      "step": 7168
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "eval_bleu": 0.564904310555124,
+      "eval_ce_clean_loss": 0.020537127653307626,
+      "eval_ce_pred_loss": 1.7842765613151257,
+      "eval_flow_consistency_loss": 0.17003079746831964,
+      "eval_flow_mse_loss": 0.9064064752826813,
+      "eval_loss": 2.2609525786788223,
+      "eval_runtime": 222.6102,
+      "eval_samples_per_second": 134.765,
+      "eval_steps_per_second": 2.107,
+      "flow/cos_sim": 0.7994557322977957,
+      "flow/improvement_ratio": 0.9958853139551972,
+      "flow/mag_ratio_mean": 0.7681991188510903,
+      "flow/mag_ratio_std": 0.12335055292860023,
+      "step": 7168
+    },
+    {
+      "epoch": 0.353050186394294,
+      "grad_norm": 0.5962383151054382,
+      "learning_rate": 7.638108666591397e-05,
+      "loss": 2.3131587505340576,
+      "step": 8192
+    },
+    {
+      "epoch": 0.353050186394294,
+      "eval_bleu": 0.5698637066625777,
+      "eval_ce_clean_loss": 0.015508384108225673,
+      "eval_ce_pred_loss": 1.7501515601235411,
+      "eval_flow_consistency_loss": 0.15371561949568263,
+      "eval_flow_mse_loss": 0.9227519265370074,
+      "eval_loss": 2.240224189595627,
+      "flow/cos_sim": 0.809604713911695,
+      "flow/improvement_ratio": 0.9950154384316158,
+      "flow/mag_ratio_mean": 0.7853530231061012,
+      "flow/mag_ratio_std": 0.1324605667419525,
+      "step": 8192
+    },
+    {
+      "epoch": 0.353050186394294,
+      "eval_bleu": 0.5698637066625777,
+      "eval_ce_clean_loss": 0.015508384108225673,
+      "eval_ce_pred_loss": 1.7501515601235411,
+      "eval_flow_consistency_loss": 0.15371561949568263,
+      "eval_flow_mse_loss": 0.9227519265370074,
+      "eval_loss": 2.240224189595627,
+      "eval_runtime": 221.2743,
+      "eval_samples_per_second": 135.578,
+      "eval_steps_per_second": 2.12,
+      "flow/cos_sim": 0.809604713911695,
+      "flow/improvement_ratio": 0.9950154384316158,
+      "flow/mag_ratio_mean": 0.7853530231061012,
+      "flow/mag_ratio_std": 0.1324605667419525,
+      "step": 8192
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "grad_norm": 0.9759780764579773,
+      "learning_rate": 6.997172522088177e-05,
+      "loss": 2.277475118637085,
+      "step": 9216
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "eval_bleu": 0.5782382780348954,
+      "eval_ce_clean_loss": 0.012276997226776916,
+      "eval_ce_pred_loss": 1.6780265332030844,
+      "eval_flow_consistency_loss": 0.14418257795163056,
+      "eval_flow_mse_loss": 0.89588076715022,
+      "eval_loss": 2.1548676056140015,
+      "flow/cos_sim": 0.8200861430371494,
+      "flow/improvement_ratio": 0.9951881179169043,
+      "flow/mag_ratio_mean": 0.7937601523867039,
+      "flow/mag_ratio_std": 0.1297300179629946,
+      "step": 9216
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "eval_bleu": 0.5782382780348954,
+      "eval_ce_clean_loss": 0.012276997226776916,
+      "eval_ce_pred_loss": 1.6780265332030844,
+      "eval_flow_consistency_loss": 0.14418257795163056,
+      "eval_flow_mse_loss": 0.89588076715022,
+      "eval_loss": 2.1548676056140015,
+      "eval_runtime": 222.088,
+      "eval_samples_per_second": 135.082,
+      "eval_steps_per_second": 2.112,
+      "flow/cos_sim": 0.8200861430371494,
+      "flow/improvement_ratio": 0.9951881179169043,
+      "flow/mag_ratio_mean": 0.7937601523867039,
+      "flow/mag_ratio_std": 0.1297300179629946,
+      "step": 9216
+    }
+  ],
+  "logging_steps": 1024,
+  "max_steps": 23204,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 1024,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 64,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoints-v5.11-b/checkpoint-9216/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8469bbc81a2ba0be2f5b44007faafd15c75615abe30f4f4e56171816d31caa5b
+size 5137