Upload folder using huggingface_hub

.gitattributes

@@ -64,3 +64,4 @@ checkpoints-v5.13-b/checkpoint-10240/eval_state.json filter=lfs diff=lfs merge=l
 checkpoints-v4.6+/checkpoint-13312/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v4.6+/checkpoint-21504/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v5.13-c/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs -text
+checkpoints-v5.14/checkpoint-10240/eval_state.json filter=lfs diff=lfs merge=lfs -text

checkpoints-v5.14/checkpoint-10240/ema.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e082830610e4245140df8200f3d4174a0dffd192e9cf2c4be2a980375de0e939
+size 54599376

checkpoints-v5.14/checkpoint-10240/eval_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e8cdde21410a492b65a2ba0c171dbe89057476e658aa2a3ff26e1ca634a1f04f
+size 61017793

checkpoints-v5.14/checkpoint-10240/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9f0ae1cfc8c2e1d44299eddd389440fbfce04e39b3510a5b5b136e130f388ec7
+size 54599408

checkpoints-v5.14/checkpoint-10240/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c555c39c7633b19bb3e6b1c1c30b6ceba99d461ae2725e3f204f071c506010b
+size 76550347

checkpoints-v5.14/checkpoint-10240/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b9bbc5571061a9947e56b5967e639907cac40378f4dc4401afa84be8834680ad
+size 14645

checkpoints-v5.14/checkpoint-10240/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ba4f519d8b83feceb2161ce743d105543b90d9de6084f72225671250185eba01
+size 1383

checkpoints-v5.14/checkpoint-10240/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:012f7b23a26c3659b3d642258b23d2a9db5c112a8101c3ff10a13ac93f47a91e
+size 1465

checkpoints-v5.14/checkpoint-10240/trainer_state.json

@@ -0,0 +1,394 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.44131273299286744,
+  "eval_steps": 1024,
+  "global_step": 10240,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.04413127329928675,
+      "grad_norm": 2.3938751220703125,
+      "learning_rate": 9.990234375e-05,
+      "loss": 7.639987945556641,
+      "step": 1024
+    },
+    {
+      "epoch": 0.04413127329928675,
+      "eval_bleu": 0.3665367955679541,
+      "eval_ce_clean_loss": 2.1982549197638215,
+      "eval_ce_pred_loss": 4.012731915597977,
+      "eval_flow_mse_loss": 0.6777123596622492,
+      "eval_loss": 3.277240484253938,
+      "flow/cos_sim": 0.5957727816059137,
+      "flow/improvement_ratio": 0.9601070273405453,
+      "flow/mag_ratio_mean": 0.5324122500317946,
+      "flow/mag_ratio_std": 0.14340894446888966,
+      "step": 1024
+    },
+    {
+      "epoch": 0.04413127329928675,
+      "eval_bleu": 0.3665367955679541,
+      "eval_ce_clean_loss": 2.1982549197638215,
+      "eval_ce_pred_loss": 4.012731915597977,
+      "eval_flow_mse_loss": 0.6777123596622492,
+      "eval_loss": 3.277240484253938,
+      "eval_runtime": 210.3423,
+      "eval_samples_per_second": 142.625,
+      "eval_steps_per_second": 2.23,
+      "flow/cos_sim": 0.5957727816059137,
+      "flow/improvement_ratio": 0.9601070273405453,
+      "flow/mag_ratio_mean": 0.5324122500317946,
+      "flow/mag_ratio_std": 0.14340894446888966,
+      "step": 1024
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "grad_norm": 2.127551794052124,
+      "learning_rate": 9.9476028157316e-05,
+      "loss": 1.9875907897949219,
+      "step": 2048
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "eval_bleu": 0.47239301802941436,
+      "eval_ce_clean_loss": 0.4944258190548496,
+      "eval_ce_pred_loss": 2.776211817635656,
+      "eval_flow_mse_loss": 0.4178365491854865,
+      "eval_loss": 1.1898835559389485,
+      "flow/cos_sim": 0.7644255845023117,
+      "flow/improvement_ratio": 0.9770590040221143,
+      "flow/mag_ratio_mean": 0.7315180154243258,
+      "flow/mag_ratio_std": 0.20312415069735634,
+      "step": 2048
+    },
+    {
+      "epoch": 0.0882625465985735,
+      "eval_bleu": 0.47239301802941436,
+      "eval_ce_clean_loss": 0.4944258190548496,
+      "eval_ce_pred_loss": 2.776211817635656,
+      "eval_flow_mse_loss": 0.4178365491854865,
+      "eval_loss": 1.1898835559389485,
+      "eval_runtime": 205.6011,
+      "eval_samples_per_second": 145.914,
+      "eval_steps_per_second": 2.281,
+      "flow/cos_sim": 0.7644255845023117,
+      "flow/improvement_ratio": 0.9770590040221143,
+      "flow/mag_ratio_mean": 0.7315180154243258,
+      "flow/mag_ratio_std": 0.20312415069735634,
+      "step": 2048
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "grad_norm": 0.5170137286186218,
+      "learning_rate": 9.791307026072513e-05,
+      "loss": 0.8957924842834473,
+      "step": 3072
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "eval_bleu": 0.5217005787375273,
+      "eval_ce_clean_loss": 0.12680107509212962,
+      "eval_ce_pred_loss": 2.1341091328338266,
+      "eval_flow_mse_loss": 0.3295880077617255,
+      "eval_loss": 0.669800000404244,
+      "flow/cos_sim": 0.8211929453715586,
+      "flow/improvement_ratio": 0.9835899890358768,
+      "flow/mag_ratio_mean": 0.7927031459838851,
+      "flow/mag_ratio_std": 0.19854979916040832,
+      "step": 3072
+    },
+    {
+      "epoch": 0.13239381989786023,
+      "eval_bleu": 0.5217005787375273,
+      "eval_ce_clean_loss": 0.12680107509212962,
+      "eval_ce_pred_loss": 2.1341091328338266,
+      "eval_flow_mse_loss": 0.3295880077617255,
+      "eval_loss": 0.669800000404244,
+      "eval_runtime": 207.9355,
+      "eval_samples_per_second": 144.276,
+      "eval_steps_per_second": 2.256,
+      "flow/cos_sim": 0.8211929453715586,
+      "flow/improvement_ratio": 0.9835899890358768,
+      "flow/mag_ratio_mean": 0.7927031459838851,
+      "flow/mag_ratio_std": 0.19854979916040832,
+      "step": 3072
+    },
+    {
+      "epoch": 0.176525093197147,
+      "grad_norm": 0.926017701625824,
+      "learning_rate": 9.53439476074686e-05,
+      "loss": 0.6218850016593933,
+      "step": 4096
+    },
+    {
+      "epoch": 0.176525093197147,
+      "eval_bleu": 0.543675748301631,
+      "eval_ce_clean_loss": 0.05681392484938285,
+      "eval_ce_pred_loss": 1.8018918128918484,
+      "eval_flow_mse_loss": 0.3036336704040133,
+      "eval_loss": 0.5406367787038848,
+      "flow/cos_sim": 0.8355540915338724,
+      "flow/improvement_ratio": 0.9839625680116194,
+      "flow/mag_ratio_mean": 0.8140800381774334,
+      "flow/mag_ratio_std": 0.20049285866431335,
+      "step": 4096
+    },
+    {
+      "epoch": 0.176525093197147,
+      "eval_bleu": 0.543675748301631,
+      "eval_ce_clean_loss": 0.05681392484938285,
+      "eval_ce_pred_loss": 1.8018918128918484,
+      "eval_flow_mse_loss": 0.3036336704040133,
+      "eval_loss": 0.5406367787038848,
+      "eval_runtime": 208.0044,
+      "eval_samples_per_second": 144.228,
+      "eval_steps_per_second": 2.255,
+      "flow/cos_sim": 0.8355540915338724,
+      "flow/improvement_ratio": 0.9839625680116194,
+      "flow/mag_ratio_mean": 0.8140800381774334,
+      "flow/mag_ratio_std": 0.20049285866431335,
+      "step": 4096
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "grad_norm": 0.8565777540206909,
+      "learning_rate": 9.182261125213742e-05,
+      "loss": 0.533421516418457,
+      "step": 5120
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "eval_bleu": 0.5544799404246258,
+      "eval_ce_clean_loss": 0.03229137011237744,
+      "eval_ce_pred_loss": 1.5962520919120642,
+      "eval_flow_mse_loss": 0.2900454838202198,
+      "eval_loss": 0.481962066596505,
+      "flow/cos_sim": 0.8430635610114791,
+      "flow/improvement_ratio": 0.9848958956661509,
+      "flow/mag_ratio_mean": 0.8249702423111971,
+      "flow/mag_ratio_std": 0.18978321196427986,
+      "step": 5120
+    },
+    {
+      "epoch": 0.22065636649643372,
+      "eval_bleu": 0.5544799404246258,
+      "eval_ce_clean_loss": 0.03229137011237744,
+      "eval_ce_pred_loss": 1.5962520919120642,
+      "eval_flow_mse_loss": 0.2900454838202198,
+      "eval_loss": 0.481962066596505,
+      "eval_runtime": 209.2596,
+      "eval_samples_per_second": 143.363,
+      "eval_steps_per_second": 2.241,
+      "flow/cos_sim": 0.8430635610114791,
+      "flow/improvement_ratio": 0.9848958956661509,
+      "flow/mag_ratio_mean": 0.8249702423111971,
+      "flow/mag_ratio_std": 0.18978321196427986,
+      "step": 5120
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "grad_norm": 0.459722101688385,
+      "learning_rate": 8.742770483354739e-05,
+      "loss": 0.4872491657733917,
+      "step": 6144
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "eval_bleu": 0.5527687698410321,
+      "eval_ce_clean_loss": 0.020704815237681622,
+      "eval_ce_pred_loss": 1.5302676085724252,
+      "eval_flow_mse_loss": 0.27337305222365904,
+      "eval_loss": 0.4471046316471181,
+      "flow/cos_sim": 0.8521691379008263,
+      "flow/improvement_ratio": 0.9867461412700255,
+      "flow/mag_ratio_mean": 0.8359249935729671,
+      "flow/mag_ratio_std": 0.1747948888903742,
+      "step": 6144
+    },
+    {
+      "epoch": 0.26478763979572045,
+      "eval_bleu": 0.5527687698410321,
+      "eval_ce_clean_loss": 0.020704815237681622,
+      "eval_ce_pred_loss": 1.5302676085724252,
+      "eval_flow_mse_loss": 0.27337305222365904,
+      "eval_loss": 0.4471046316471181,
+      "eval_runtime": 209.107,
+      "eval_samples_per_second": 143.467,
+      "eval_steps_per_second": 2.243,
+      "flow/cos_sim": 0.8521691379008263,
+      "flow/improvement_ratio": 0.9867461412700255,
+      "flow/mag_ratio_mean": 0.8359249935729671,
+      "flow/mag_ratio_std": 0.1747948888903742,
+      "step": 6144
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "grad_norm": 0.5934994220733643,
+      "learning_rate": 8.224294338515429e-05,
+      "loss": 0.45292773842811584,
+      "step": 7168
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "eval_bleu": 0.5556870699184919,
+      "eval_ce_clean_loss": 0.014376114403554944,
+      "eval_ce_pred_loss": 1.4703641462681898,
+      "eval_flow_mse_loss": 0.25588471872974305,
+      "eval_loss": 0.4172972510618442,
+      "flow/cos_sim": 0.8633063356759452,
+      "flow/improvement_ratio": 0.9885822310884879,
+      "flow/mag_ratio_mean": 0.8422111430401995,
+      "flow/mag_ratio_std": 0.16398038163876483,
+      "step": 7168
+    },
+    {
+      "epoch": 0.30891891309500724,
+      "eval_bleu": 0.5556870699184919,
+      "eval_ce_clean_loss": 0.014376114403554944,
+      "eval_ce_pred_loss": 1.4703641462681898,
+      "eval_flow_mse_loss": 0.25588471872974305,
+      "eval_loss": 0.4172972510618442,
+      "eval_runtime": 209.704,
+      "eval_samples_per_second": 143.059,
+      "eval_steps_per_second": 2.236,
+      "flow/cos_sim": 0.8633063356759452,
+      "flow/improvement_ratio": 0.9885822310884879,
+      "flow/mag_ratio_mean": 0.8422111430401995,
+      "flow/mag_ratio_std": 0.16398038163876483,
+      "step": 7168
+    },
+    {
+      "epoch": 0.353050186394294,
+      "grad_norm": 0.3148420453071594,
+      "learning_rate": 7.638710244802891e-05,
+      "loss": 0.4292431175708771,
+      "step": 8192
+    },
+    {
+      "epoch": 0.353050186394294,
+      "eval_bleu": 0.5566824370986413,
+      "eval_ce_clean_loss": 0.010665512040082707,
+      "eval_ce_pred_loss": 1.437801378749327,
+      "eval_flow_mse_loss": 0.24223278170582582,
+      "eval_loss": 0.39667843399779884,
+      "flow/cos_sim": 0.8718424904575226,
+      "flow/improvement_ratio": 0.9907243636879586,
+      "flow/mag_ratio_mean": 0.8495659919689967,
+      "flow/mag_ratio_std": 0.1551145709780996,
+      "step": 8192
+    },
+    {
+      "epoch": 0.353050186394294,
+      "eval_bleu": 0.5566824370986413,
+      "eval_ce_clean_loss": 0.010665512040082707,
+      "eval_ce_pred_loss": 1.437801378749327,
+      "eval_flow_mse_loss": 0.24223278170582582,
+      "eval_loss": 0.39667843399779884,
+      "eval_runtime": 206.6493,
+      "eval_samples_per_second": 145.173,
+      "eval_steps_per_second": 2.27,
+      "flow/cos_sim": 0.8718424904575226,
+      "flow/improvement_ratio": 0.9907243636879586,
+      "flow/mag_ratio_mean": 0.8495659919689967,
+      "flow/mag_ratio_std": 0.1551145709780996,
+      "step": 8192
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "grad_norm": 0.5599232912063599,
+      "learning_rate": 6.997172522088177e-05,
+      "loss": 0.41261157393455505,
+      "step": 9216
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "eval_bleu": 0.5663410724389681,
+      "eval_ce_clean_loss": 0.008277329295051537,
+      "eval_ce_pred_loss": 1.3597534239165057,
+      "eval_flow_mse_loss": 0.2412289542747713,
+      "eval_loss": 0.38548162909967304,
+      "flow/cos_sim": 0.8725989290646144,
+      "flow/improvement_ratio": 0.9909563410256718,
+      "flow/mag_ratio_mean": 0.8561376066350225,
+      "flow/mag_ratio_std": 0.15248549610439902,
+      "step": 9216
+    },
+    {
+      "epoch": 0.3971814596935807,
+      "eval_bleu": 0.5663410724389681,
+      "eval_ce_clean_loss": 0.008277329295051537,
+      "eval_ce_pred_loss": 1.3597534239165057,
+      "eval_flow_mse_loss": 0.2412289542747713,
+      "eval_loss": 0.38548162909967304,
+      "eval_runtime": 208.3126,
+      "eval_samples_per_second": 144.014,
+      "eval_steps_per_second": 2.251,
+      "flow/cos_sim": 0.8725989290646144,
+      "flow/improvement_ratio": 0.9909563410256718,
+      "flow/mag_ratio_mean": 0.8561376066350225,
+      "flow/mag_ratio_std": 0.15248549610439902,
+      "step": 9216
+    },
+    {
+      "epoch": 0.44131273299286744,
+      "grad_norm": 0.5740413069725037,
+      "learning_rate": 6.314377890922702e-05,
+      "loss": 0.4009917378425598,
+      "step": 10240
+    },
+    {
+      "epoch": 0.44131273299286744,
+      "eval_bleu": 0.5765275121149777,
+      "eval_ce_clean_loss": 0.0067116839581353865,
+      "eval_ce_pred_loss": 1.3255969242754775,
+      "eval_flow_mse_loss": 0.23518445449216024,
+      "eval_loss": 0.37445583431197127,
+      "flow/cos_sim": 0.8770843943807362,
+      "flow/improvement_ratio": 0.9924178827546044,
+      "flow/mag_ratio_mean": 0.855633486181434,
+      "flow/mag_ratio_std": 0.14714058946127068,
+      "step": 10240
+    },
+    {
+      "epoch": 0.44131273299286744,
+      "eval_bleu": 0.5765275121149777,
+      "eval_ce_clean_loss": 0.0067116839581353865,
+      "eval_ce_pred_loss": 1.3255969242754775,
+      "eval_flow_mse_loss": 0.23518445449216024,
+      "eval_loss": 0.37445583431197127,
+      "eval_runtime": 208.5305,
+      "eval_samples_per_second": 143.864,
+      "eval_steps_per_second": 2.249,
+      "flow/cos_sim": 0.8770843943807362,
+      "flow/improvement_ratio": 0.9924178827546044,
+      "flow/mag_ratio_mean": 0.855633486181434,
+      "flow/mag_ratio_std": 0.14714058946127068,
+      "step": 10240
+    }
+  ],
+  "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-v5.14/checkpoint-10240/training_args.bin

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