Upload folder using huggingface_hub

.gitattributes

@@ -48,3 +48,4 @@ checkpoints-v5.7/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs
 checkpoints-v5.8/checkpoint-10240/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v4.5/checkpoint-7168/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v4.5/checkpoint-15360/eval_state.json filter=lfs diff=lfs merge=lfs -text
+checkpoints-v5.9/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs -text

checkpoints-v5.9/checkpoint-11264/ema.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d50c375b4adfdb6701b14ccf091d1af64e58a8bf6bc3b23d0b0f0572c48d2c88
+size 54599592

checkpoints-v5.9/checkpoint-11264/eval_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b2a5c0a3ba0a45dd9a72c8140e6ddaf369fd02a427295653585d7c8b8422b3de
+size 59623194

checkpoints-v5.9/checkpoint-11264/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b6d9dd468e268146a81f58330f565165ba042cb8cd627c90c26564638bcaf8f
+size 54599624

checkpoints-v5.9/checkpoint-11264/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3a6467fedb0537962ac0470155c14409beedb357936bfeb0262ddbfa5a31e88
+size 76550347

checkpoints-v5.9/checkpoint-11264/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3445e2029753be8bf32790ad2df6092dc63f2028e7603a9903f650b4f19cbab4
+size 14645

checkpoints-v5.9/checkpoint-11264/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1763c6e6bb9d17ce1be6965092741353bef140a65f63acbf2d5abb2b85825bac
+size 1383

checkpoints-v5.9/checkpoint-11264/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19ad511dc53ec97f3489767f5cff37a551a58fb728bc82956d6ea6fdb5514518
+size 1465

checkpoints-v5.9/checkpoint-11264/trainer_state.json

@@ -0,0 +1,430 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.4854440062921542,
+  "eval_steps": 1024,
+  "global_step": 11264,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04413127329928675,
+      "grad_norm": 15.239693641662598,
+      "learning_rate": 9.990234375e-05,
+      "loss": 13.825879096984863,
+      "step": 1024
+    },
+    {
+      "epoch": 0.04413127329928675,
+      "eval_bleu": 0.03856809374192324,
+      "eval_ce_clean_loss": 4.292109883924537,
+      "eval_ce_pred_loss": 7.204727572418734,
+      "eval_flow_mse_loss": 0.6720071626878751,
+      "eval_loss": 10.00742630257027,
+      "flow/cos_sim": 0.6058795800341218,
+      "flow/improvement_ratio": 0.9756911718514937,
+      "flow/mag_ratio_mean": 0.5525405872096893,
+      "flow/mag_ratio_std": 0.1515131940814986,
+      "step": 1024
+    },
+    {
+      "epoch": 0.04413127329928675,
+      "eval_bleu": 0.03856809374192324,
+      "eval_ce_clean_loss": 4.292109883924537,
+      "eval_ce_pred_loss": 7.204727572418734,
+      "eval_flow_mse_loss": 0.6720071626878751,
+      "eval_loss": 10.00742630257027,
+      "eval_runtime": 196.3774,
+      "eval_samples_per_second": 152.767,
+      "eval_steps_per_second": 2.388,
+      "flow/cos_sim": 0.6058795800341218,
+      "flow/improvement_ratio": 0.9756911718514937,
+      "flow/mag_ratio_mean": 0.5525405872096893,
+      "flow/mag_ratio_std": 0.1515131940814986,
+      "step": 1024
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "grad_norm": 14.527959823608398,
+      "learning_rate": 9.947705025097448e-05,
+      "loss": 8.282852172851562,
+      "step": 2048
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "eval_bleu": 0.04835653057106903,
+      "eval_ce_clean_loss": 2.0321871253218986,
+      "eval_ce_pred_loss": 6.679203874266732,
+      "eval_flow_mse_loss": 0.34213011389348047,
+      "eval_loss": 7.049759871924102,
+      "flow/cos_sim": 0.8224498067837535,
+      "flow/improvement_ratio": 0.9910931860460147,
+      "flow/mag_ratio_mean": 0.7947272139825801,
+      "flow/mag_ratio_std": 0.1784994681951588,
+      "step": 2048
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "eval_bleu": 0.04835653057106903,
+      "eval_ce_clean_loss": 2.0321871253218986,
+      "eval_ce_pred_loss": 6.679203874266732,
+      "eval_flow_mse_loss": 0.34213011389348047,
+      "eval_loss": 7.049759871924102,
+      "eval_runtime": 190.6373,
+      "eval_samples_per_second": 157.367,
+      "eval_steps_per_second": 2.46,
+      "flow/cos_sim": 0.8224498067837535,
+      "flow/improvement_ratio": 0.9910931860460147,
+      "flow/mag_ratio_mean": 0.7947272139825801,
+      "flow/mag_ratio_std": 0.1784994681951588,
+      "step": 2048
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "grad_norm": 18.57774543762207,
+      "learning_rate": 9.791711775587888e-05,
+      "loss": 6.476203441619873,
+      "step": 3072
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "eval_bleu": 0.046949984747013486,
+      "eval_ce_clean_loss": 0.9904567614547225,
+      "eval_ce_pred_loss": 6.729919616601615,
+      "eval_flow_mse_loss": 0.2399635909081522,
+      "eval_loss": 5.941364000601046,
+      "flow/cos_sim": 0.8910990282416598,
+      "flow/improvement_ratio": 0.9930302517246336,
+      "flow/mag_ratio_mean": 0.879739438038645,
+      "flow/mag_ratio_std": 0.15084654318371307,
+      "step": 3072
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "eval_bleu": 0.046949984747013486,
+      "eval_ce_clean_loss": 0.9904567614547225,
+      "eval_ce_pred_loss": 6.729919616601615,
+      "eval_flow_mse_loss": 0.2399635909081522,
+      "eval_loss": 5.941364000601046,
+      "eval_runtime": 193.1899,
+      "eval_samples_per_second": 155.288,
+      "eval_steps_per_second": 2.428,
+      "flow/cos_sim": 0.8910990282416598,
+      "flow/improvement_ratio": 0.9930302517246336,
+      "flow/mag_ratio_mean": 0.879739438038645,
+      "flow/mag_ratio_std": 0.15084654318371307,
+      "step": 3072
+    },
+    {
+      "epoch": 0.176525093197147,
+      "grad_norm": 17.850786209106445,
+      "learning_rate": 9.534991440649608e-05,
+      "loss": 5.697450637817383,
+      "step": 4096
+    },
+    {
+      "epoch": 0.176525093197147,
+      "eval_bleu": 0.0394065103208499,
+      "eval_ce_clean_loss": 0.47789636076386294,
+      "eval_ce_pred_loss": 6.793904790491946,
+      "eval_flow_mse_loss": 0.19059975353131162,
+      "eval_loss": 5.424229394144087,
+      "flow/cos_sim": 0.9323588246221481,
+      "flow/improvement_ratio": 0.9941833016714816,
+      "flow/mag_ratio_mean": 0.9235222680228097,
+      "flow/mag_ratio_std": 0.12112796954762961,
+      "step": 4096
+    },
+    {
+      "epoch": 0.176525093197147,
+      "eval_bleu": 0.0394065103208499,
+      "eval_ce_clean_loss": 0.47789636076386294,
+      "eval_ce_pred_loss": 6.793904790491946,
+      "eval_flow_mse_loss": 0.19059975353131162,
+      "eval_loss": 5.424229394144087,
+      "eval_runtime": 194.4713,
+      "eval_samples_per_second": 154.264,
+      "eval_steps_per_second": 2.412,
+      "flow/cos_sim": 0.9323588246221481,
+      "flow/improvement_ratio": 0.9941833016714816,
+      "flow/mag_ratio_mean": 0.9235222680228097,
+      "flow/mag_ratio_std": 0.12112796954762961,
+      "step": 4096
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "grad_norm": 16.39472007751465,
+      "learning_rate": 9.183037205346935e-05,
+      "loss": 5.361335277557373,
+      "step": 5120
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "eval_bleu": 0.03285729536649342,
+      "eval_ce_clean_loss": 0.24329622429825348,
+      "eval_ce_pred_loss": 6.871835609997259,
+      "eval_flow_mse_loss": 0.1694319029606736,
+      "eval_loss": 5.223012985196958,
+      "flow/cos_sim": 0.9513821737852686,
+      "flow/improvement_ratio": 0.9946333313547472,
+      "flow/mag_ratio_mean": 0.9428434033892048,
+      "flow/mag_ratio_std": 0.09642866442897427,
+      "step": 5120
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "eval_bleu": 0.03285729536649342,
+      "eval_ce_clean_loss": 0.24329622429825348,
+      "eval_ce_pred_loss": 6.871835609997259,
+      "eval_flow_mse_loss": 0.1694319029606736,
+      "eval_loss": 5.223012985196958,
+      "eval_runtime": 194.2427,
+      "eval_samples_per_second": 154.446,
+      "eval_steps_per_second": 2.415,
+      "flow/cos_sim": 0.9513821737852686,
+      "flow/improvement_ratio": 0.9946333313547472,
+      "flow/mag_ratio_mean": 0.9428434033892048,
+      "flow/mag_ratio_std": 0.09642866442897427,
+      "step": 5120
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "grad_norm": 13.939663887023926,
+      "learning_rate": 8.74324003722993e-05,
+      "loss": 5.212263107299805,
+      "step": 6144
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "eval_bleu": 0.027256566689703093,
+      "eval_ce_clean_loss": 0.1259780960646011,
+      "eval_ce_pred_loss": 6.923184791353465,
+      "eval_flow_mse_loss": 0.15653546359429735,
+      "eval_loss": 5.128742840244318,
+      "flow/cos_sim": 0.9604966702745922,
+      "flow/improvement_ratio": 0.9943097961991072,
+      "flow/mag_ratio_mean": 0.9571583648480332,
+      "flow/mag_ratio_std": 0.08279879235509616,
+      "step": 6144
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "eval_bleu": 0.027256566689703093,
+      "eval_ce_clean_loss": 0.1259780960646011,
+      "eval_ce_pred_loss": 6.923184791353465,
+      "eval_flow_mse_loss": 0.15653546359429735,
+      "eval_loss": 5.128742840244318,
+      "eval_runtime": 193.2759,
+      "eval_samples_per_second": 155.219,
+      "eval_steps_per_second": 2.427,
+      "flow/cos_sim": 0.9604966702745922,
+      "flow/improvement_ratio": 0.9943097961991072,
+      "flow/mag_ratio_mean": 0.9571583648480332,
+      "flow/mag_ratio_std": 0.08279879235509616,
+      "step": 6144
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "grad_norm": 16.255971908569336,
+      "learning_rate": 8.225917891725653e-05,
+      "loss": 5.156039714813232,
+      "step": 7168
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "eval_bleu": 0.022155517814971083,
+      "eval_ce_clean_loss": 0.07260342659567719,
+      "eval_ce_pred_loss": 6.978886835102333,
+      "eval_flow_mse_loss": 0.14691995797571597,
+      "eval_loss": 5.104744104942533,
+      "flow/cos_sim": 0.9672264456748962,
+      "flow/improvement_ratio": 0.9955292942681546,
+      "flow/mag_ratio_mean": 0.9630893364644,
+      "flow/mag_ratio_std": 0.07540169008759293,
+      "step": 7168
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "eval_bleu": 0.022155517814971083,
+      "eval_ce_clean_loss": 0.07260342659567719,
+      "eval_ce_pred_loss": 6.978886835102333,
+      "eval_flow_mse_loss": 0.14691995797571597,
+      "eval_loss": 5.104744104942533,
+      "eval_runtime": 195.9924,
+      "eval_samples_per_second": 153.067,
+      "eval_steps_per_second": 2.393,
+      "flow/cos_sim": 0.9672264456748962,
+      "flow/improvement_ratio": 0.9955292942681546,
+      "flow/mag_ratio_mean": 0.9630893364644,
+      "flow/mag_ratio_std": 0.07540169008759293,
+      "step": 7168
+    },
+    {
+      "epoch": 0.353050186394294,
+      "grad_norm": 8.994356155395508,
+      "learning_rate": 7.639913242399507e-05,
+      "loss": 5.134494781494141,
+      "step": 8192
+    },
+    {
+      "epoch": 0.353050186394294,
+      "eval_bleu": 0.019220227732395335,
+      "eval_ce_clean_loss": 0.04499363649819197,
+      "eval_ce_pred_loss": 7.006692920920691,
+      "eval_flow_mse_loss": 0.14029116780836698,
+      "eval_loss": 5.089969773282374,
+      "flow/cos_sim": 0.9713161822829419,
+      "flow/improvement_ratio": 0.9945664123685629,
+      "flow/mag_ratio_mean": 0.9694278030507346,
+      "flow/mag_ratio_std": 0.06895820532780467,
+      "step": 8192
+    },
+    {
+      "epoch": 0.353050186394294,
+      "eval_bleu": 0.019220227732395335,
+      "eval_ce_clean_loss": 0.04499363649819197,
+      "eval_ce_pred_loss": 7.006692920920691,
+      "eval_flow_mse_loss": 0.14029116780836698,
+      "eval_loss": 5.089969773282374,
+      "eval_runtime": 194.1667,
+      "eval_samples_per_second": 154.506,
+      "eval_steps_per_second": 2.415,
+      "flow/cos_sim": 0.9713161822829419,
+      "flow/improvement_ratio": 0.9945664123685629,
+      "flow/mag_ratio_mean": 0.9694278030507346,
+      "flow/mag_ratio_std": 0.06895820532780467,
+      "step": 8192
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "grad_norm": 9.113734245300293,
+      "learning_rate": 6.998470950469718e-05,
+      "loss": 5.114821910858154,
+      "step": 9216
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "eval_bleu": 0.01583632669902971,
+      "eval_ce_clean_loss": 0.03044994743163588,
+      "eval_ce_pred_loss": 7.017744996654454,
+      "eval_flow_mse_loss": 0.1349480300824017,
+      "eval_loss": 5.077819404317371,
+      "flow/cos_sim": 0.9751363802057847,
+      "flow/improvement_ratio": 0.9953071037843537,
+      "flow/mag_ratio_mean": 0.969198439929531,
+      "flow/mag_ratio_std": 0.06459524200510369,
+      "step": 9216
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "eval_bleu": 0.01583632669902971,
+      "eval_ce_clean_loss": 0.03044994743163588,
+      "eval_ce_pred_loss": 7.017744996654454,
+      "eval_flow_mse_loss": 0.1349480300824017,
+      "eval_loss": 5.077819404317371,
+      "eval_runtime": 192.9017,
+      "eval_samples_per_second": 155.52,
+      "eval_steps_per_second": 2.431,
+      "flow/cos_sim": 0.9751363802057847,
+      "flow/improvement_ratio": 0.9953071037843537,
+      "flow/mag_ratio_mean": 0.969198439929531,
+      "flow/mag_ratio_std": 0.06459524200510369,
+      "step": 9216
+    },
+    {
+      "epoch": 0.44131273299286744,
+      "grad_norm": 7.316830158233643,
+      "learning_rate": 6.315061173955019e-05,
+      "loss": 5.1045637130737305,
+      "step": 10240
+    },
+    {
+      "epoch": 0.44131273299286744,
+      "eval_bleu": 0.01353516095055143,
+      "eval_ce_clean_loss": 0.021886831322220215,
+      "eval_ce_pred_loss": 7.025449516930814,
+      "eval_flow_mse_loss": 0.1312034219694036,
+      "eval_loss": 5.070904863922835,
+      "flow/cos_sim": 0.9770022788281634,
+      "flow/improvement_ratio": 0.9958924522786252,
+      "flow/mag_ratio_mean": 0.974764855050329,
+      "flow/mag_ratio_std": 0.06171431206563897,
+      "step": 10240
+    },
+    {
+      "epoch": 0.44131273299286744,
+      "eval_bleu": 0.01353516095055143,
+      "eval_ce_clean_loss": 0.021886831322220215,
+      "eval_ce_pred_loss": 7.025449516930814,
+      "eval_flow_mse_loss": 0.1312034219694036,
+      "eval_loss": 5.070904863922835,
+      "eval_runtime": 194.3566,
+      "eval_samples_per_second": 154.355,
+      "eval_steps_per_second": 2.413,
+      "flow/cos_sim": 0.9770022788281634,
+      "flow/improvement_ratio": 0.9958924522786252,
+      "flow/mag_ratio_mean": 0.974764855050329,
+      "flow/mag_ratio_std": 0.06171431206563897,
+      "step": 10240
+    },
+    {
+      "epoch": 0.4854440062921542,
+      "grad_norm": 8.874271392822266,
+      "learning_rate": 5.604738390528452e-05,
+      "loss": 5.102542400360107,
+      "step": 11264
+    },
+    {
+      "epoch": 0.4854440062921542,
+      "eval_bleu": 0.012540996821431605,
+      "eval_ce_clean_loss": 0.016188185467426456,
+      "eval_ce_pred_loss": 7.027885409560539,
+      "eval_flow_mse_loss": 0.12859490318402553,
+      "eval_loss": 5.064302796239792,
+      "flow/cos_sim": 0.9777249371065005,
+      "flow/improvement_ratio": 0.9943972062200371,
+      "flow/mag_ratio_mean": 0.9753310384272512,
+      "flow/mag_ratio_std": 0.05981211423842129,
+      "step": 11264
+    },
+    {
+      "epoch": 0.4854440062921542,
+      "eval_bleu": 0.012540996821431605,
+      "eval_ce_clean_loss": 0.016188185467426456,
+      "eval_ce_pred_loss": 7.027885409560539,
+      "eval_flow_mse_loss": 0.12859490318402553,
+      "eval_loss": 5.064302796239792,
+      "eval_runtime": 193.4331,
+      "eval_samples_per_second": 155.092,
+      "eval_steps_per_second": 2.425,
+      "flow/cos_sim": 0.9777249371065005,
+      "flow/improvement_ratio": 0.9943972062200371,
+      "flow/mag_ratio_mean": 0.9753310384272512,
+      "flow/mag_ratio_std": 0.05981211423842129,
+      "step": 11264
+    }
+  ],
+  "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-v5.9/checkpoint-11264/training_args.bin

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