Upload folder using huggingface_hub

.gitattributes

@@ -40,3 +40,4 @@ checkpoints-v5.3/checkpoint-11264/eval_state.json filter=lfs diff=lfs merge=lfs
 checkpoints-v5.4/checkpoint-12288/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v5.5/checkpoint-10240/eval_state.json filter=lfs diff=lfs merge=lfs -text
 checkpoints-v5.5/checkpoint-16384/eval_state.json filter=lfs diff=lfs merge=lfs -text
+checkpoints-v4.4+/checkpoint-7168/eval_state.json filter=lfs diff=lfs merge=lfs -text

checkpoints-v4.4+/checkpoint-7168/ema.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8da5fec378c5b361daa721374f0586d4663d2d8c3d08fcaadd1abb29064c45c7
+size 38738976

checkpoints-v4.4+/checkpoint-7168/eval_state.json

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:908335c0e64254ef2c23e47c85003b3b3cb509a0ba715415cb05876bbda2daf1
+size 47930611

checkpoints-v4.4+/checkpoint-7168/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2c7bfa3a8cddeaee4a71dd4f488c21fe4167e8f30bbcff08475eb649920ba84d
+size 38739008

checkpoints-v4.4+/checkpoint-7168/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fae0320c5d1c2a637abdd69ceb85aa6eb3529c11d58a6371edff21ba4d8577c3
+size 77599691

checkpoints-v4.4+/checkpoint-7168/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df98b42eddbbe331a59c4c99414ce4ee48c424c0abd57f14e539bceba56714db
+size 14645

checkpoints-v4.4+/checkpoint-7168/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b3031d089d4b9cc720ccc9e3882646caf7e007b04ab200bfcffd3c320ef9cf6d
+size 1383

checkpoints-v4.4+/checkpoint-7168/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bf421e7a76e065aa2ac77b8c072da7efc0d488f8fa73d18cfcaf6079f63bbd32
+size 1465

checkpoints-v4.4+/checkpoint-7168/trainer_state.json

@@ -0,0 +1,489 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.3310701584222438,
+  "eval_steps": 1024,
+  "global_step": 7168,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.011823934229365849,
+      "grad_norm": 2.6977920532226562,
+      "learning_rate": 0.000498046875,
+      "loss": 26.693201065063477,
+      "step": 256
+    },
+    {
+      "epoch": 0.023647868458731697,
+      "grad_norm": 2.7128376960754395,
+      "learning_rate": 0.000998046875,
+      "loss": 9.865509033203125,
+      "step": 512
+    },
+    {
+      "epoch": 0.03547180268809755,
+      "grad_norm": 3.0180280208587646,
+      "learning_rate": 0.000999961364193767,
+      "loss": 8.319804191589355,
+      "step": 768
+    },
+    {
+      "epoch": 0.047295736917463395,
+      "grad_norm": 3.1472654342651367,
+      "learning_rate": 0.0009998448561547506,
+      "loss": 7.783344745635986,
+      "step": 1024
+    },
+    {
+      "epoch": 0.047295736917463395,
+      "eval_bleu": 0.5080180462841996,
+      "eval_ce_loss": 0.027617456312780373,
+      "eval_cos_loss": 0.4708993843156997,
+      "eval_cov_loss": 0.00034012032398867916,
+      "eval_flow_ce_loss": 2.2868199040900627,
+      "eval_loss": 7.60784471307171,
+      "eval_mean_loss": 0.0015031876431599463,
+      "eval_mse_loss": 1.3905650837236343,
+      "eval_whiten_loss": 0.1768301349796661,
+      "flow/cos_sim": 0.529100628748332,
+      "flow/improvement_ratio": 0.9663206476326947,
+      "flow/mag_ratio_mean": 0.49323171737803717,
+      "flow/mag_ratio_std": 0.1666014238375507,
+      "step": 1024
+    },
+    {
+      "epoch": 0.047295736917463395,
+      "eval_bleu": 0.5080180462841996,
+      "eval_ce_loss": 0.027617456312780373,
+      "eval_cos_loss": 0.4708993843156997,
+      "eval_cov_loss": 0.00034012032398867916,
+      "eval_flow_ce_loss": 2.2868199040900627,
+      "eval_loss": 7.60784471307171,
+      "eval_mean_loss": 0.0015031876431599463,
+      "eval_mse_loss": 1.3905650837236343,
+      "eval_runtime": 137.3378,
+      "eval_samples_per_second": 203.826,
+      "eval_steps_per_second": 3.189,
+      "eval_whiten_loss": 0.1768301349796661,
+      "flow/cos_sim": 0.529100628748332,
+      "flow/improvement_ratio": 0.9663206476326947,
+      "flow/mag_ratio_mean": 0.49323171737803717,
+      "flow/mag_ratio_std": 0.1666014238375507,
+      "step": 1024
+    },
+    {
+      "epoch": 0.05911967114682925,
+      "grad_norm": 3.362802505493164,
+      "learning_rate": 0.0009996504934357976,
+      "loss": 7.467670917510986,
+      "step": 1280
+    },
+    {
+      "epoch": 0.0709436053761951,
+      "grad_norm": 2.9492599964141846,
+      "learning_rate": 0.000999378306310396,
+      "loss": 7.205803871154785,
+      "step": 1536
+    },
+    {
+      "epoch": 0.08276753960556095,
+      "grad_norm": 3.097372055053711,
+      "learning_rate": 0.0009990283371737838,
+      "loss": 7.063663959503174,
+      "step": 1792
+    },
+    {
+      "epoch": 0.09459147383492679,
+      "grad_norm": 4.688446521759033,
+      "learning_rate": 0.0009986006405363448,
+      "loss": 6.885957717895508,
+      "step": 2048
+    },
+    {
+      "epoch": 0.09459147383492679,
+      "eval_bleu": 0.5176616016025543,
+      "eval_ce_loss": 0.0072827782079505155,
+      "eval_cos_loss": 0.3974148456228378,
+      "eval_cov_loss": 0.0002958967598470366,
+      "eval_flow_ce_loss": 2.1159514251364966,
+      "eval_loss": 6.8805486981727215,
+      "eval_mean_loss": 0.0009597052286323756,
+      "eval_mse_loss": 1.221881123166106,
+      "eval_whiten_loss": 0.15687005378339933,
+      "flow/cos_sim": 0.6025851748577536,
+      "flow/improvement_ratio": 0.9678549095648065,
+      "flow/mag_ratio_mean": 0.5845441620915992,
+      "flow/mag_ratio_std": 0.1977017904253311,
+      "step": 2048
+    },
+    {
+      "epoch": 0.09459147383492679,
+      "eval_bleu": 0.5176616016025543,
+      "eval_ce_loss": 0.0072827782079505155,
+      "eval_cos_loss": 0.3974148456228378,
+      "eval_cov_loss": 0.0002958967598470366,
+      "eval_flow_ce_loss": 2.1159514251364966,
+      "eval_loss": 6.8805486981727215,
+      "eval_mean_loss": 0.0009597052286323756,
+      "eval_mse_loss": 1.221881123166106,
+      "eval_runtime": 133.2951,
+      "eval_samples_per_second": 210.008,
+      "eval_steps_per_second": 3.286,
+      "eval_whiten_loss": 0.15687005378339933,
+      "flow/cos_sim": 0.6025851748577536,
+      "flow/improvement_ratio": 0.9678549095648065,
+      "flow/mag_ratio_mean": 0.5845441620915992,
+      "flow/mag_ratio_std": 0.1977017904253311,
+      "step": 2048
+    },
+    {
+      "epoch": 0.10641540806429264,
+      "grad_norm": 2.493529796600342,
+      "learning_rate": 0.0009980952830151192,
+      "loss": 6.838229179382324,
+      "step": 2304
+    },
+    {
+      "epoch": 0.1182393422936585,
+      "grad_norm": 2.93609356880188,
+      "learning_rate": 0.0009975123433234273,
+      "loss": 6.67982816696167,
+      "step": 2560
+    },
+    {
+      "epoch": 0.13006327652302435,
+      "grad_norm": 2.9494926929473877,
+      "learning_rate": 0.0009968519122586083,
+      "loss": 6.686809539794922,
+      "step": 2816
+    },
+    {
+      "epoch": 0.1418872107523902,
+      "grad_norm": 4.821794033050537,
+      "learning_rate": 0.0009961140926878786,
+      "loss": 6.529831409454346,
+      "step": 3072
+    },
+    {
+      "epoch": 0.1418872107523902,
+      "eval_bleu": 0.536460391570044,
+      "eval_ce_loss": 0.003726785426819059,
+      "eval_cos_loss": 0.35948846853214855,
+      "eval_cov_loss": 0.0002726104497733062,
+      "eval_flow_ce_loss": 2.061866384662994,
+      "eval_loss": 6.564142907591171,
+      "eval_mean_loss": 0.000869228759422775,
+      "eval_mse_loss": 1.1264942184975157,
+      "eval_whiten_loss": 0.1468495756523794,
+      "flow/cos_sim": 0.6405115382039928,
+      "flow/improvement_ratio": 0.9704297850393269,
+      "flow/mag_ratio_mean": 0.6134878079912979,
+      "flow/mag_ratio_std": 0.19985322746103756,
+      "step": 3072
+    },
+    {
+      "epoch": 0.1418872107523902,
+      "eval_bleu": 0.536460391570044,
+      "eval_ce_loss": 0.003726785426819059,
+      "eval_cos_loss": 0.35948846853214855,
+      "eval_cov_loss": 0.0002726104497733062,
+      "eval_flow_ce_loss": 2.061866384662994,
+      "eval_loss": 6.564142907591171,
+      "eval_mean_loss": 0.000869228759422775,
+      "eval_mse_loss": 1.1264942184975157,
+      "eval_runtime": 133.1163,
+      "eval_samples_per_second": 210.29,
+      "eval_steps_per_second": 3.29,
+      "eval_whiten_loss": 0.1468495756523794,
+      "flow/cos_sim": 0.6405115382039928,
+      "flow/improvement_ratio": 0.9704297850393269,
+      "flow/mag_ratio_mean": 0.6134878079912979,
+      "flow/mag_ratio_std": 0.19985322746103756,
+      "step": 3072
+    },
+    {
+      "epoch": 0.15371114498175603,
+      "grad_norm": 3.2502996921539307,
+      "learning_rate": 0.0009952989995323104,
+      "loss": 6.486678123474121,
+      "step": 3328
+    },
+    {
+      "epoch": 0.1655350792111219,
+      "grad_norm": 2.448288917541504,
+      "learning_rate": 0.0009944067597489302,
+      "loss": 6.419068813323975,
+      "step": 3584
+    },
+    {
+      "epoch": 0.17735901344048774,
+      "grad_norm": 2.7338614463806152,
+      "learning_rate": 0.0009934375123109454,
+      "loss": 6.377047538757324,
+      "step": 3840
+    },
+    {
+      "epoch": 0.18918294766985358,
+      "grad_norm": 2.7483572959899902,
+      "learning_rate": 0.0009923914081860971,
+      "loss": 6.377966403961182,
+      "step": 4096
+    },
+    {
+      "epoch": 0.18918294766985358,
+      "eval_bleu": 0.5550769774844325,
+      "eval_ce_loss": 0.002399685377862143,
+      "eval_cos_loss": 0.3334486482486333,
+      "eval_cov_loss": 0.0002663974104274796,
+      "eval_flow_ce_loss": 1.985243722456231,
+      "eval_loss": 6.281966650322692,
+      "eval_mean_loss": 0.0007326721327249641,
+      "eval_mse_loss": 1.0656885820436695,
+      "eval_whiten_loss": 0.14359816459760275,
+      "flow/cos_sim": 0.6665513564462531,
+      "flow/improvement_ratio": 0.9709918391486826,
+      "flow/mag_ratio_mean": 0.643914669752121,
+      "flow/mag_ratio_std": 0.20350686651250544,
+      "step": 4096
+    },
+    {
+      "epoch": 0.18918294766985358,
+      "eval_bleu": 0.5550769774844325,
+      "eval_ce_loss": 0.002399685377862143,
+      "eval_cos_loss": 0.3334486482486333,
+      "eval_cov_loss": 0.0002663974104274796,
+      "eval_flow_ce_loss": 1.985243722456231,
+      "eval_loss": 6.281966650322692,
+      "eval_mean_loss": 0.0007326721327249641,
+      "eval_mse_loss": 1.0656885820436695,
+      "eval_runtime": 133.3237,
+      "eval_samples_per_second": 209.963,
+      "eval_steps_per_second": 3.285,
+      "eval_whiten_loss": 0.14359816459760275,
+      "flow/cos_sim": 0.6665513564462531,
+      "flow/improvement_ratio": 0.9709918391486826,
+      "flow/mag_ratio_mean": 0.643914669752121,
+      "flow/mag_ratio_std": 0.20350686651250544,
+      "step": 4096
+    },
+    {
+      "epoch": 0.20100688189921945,
+      "grad_norm": 2.555222272872925,
+      "learning_rate": 0.0009912686103131475,
+      "loss": 6.325072288513184,
+      "step": 4352
+    },
+    {
+      "epoch": 0.2128308161285853,
+      "grad_norm": 2.890451431274414,
+      "learning_rate": 0.0009900692935764985,
+      "loss": 6.253594875335693,
+      "step": 4608
+    },
+    {
+      "epoch": 0.22465475035795113,
+      "grad_norm": 2.830192804336548,
+      "learning_rate": 0.0009887936447789544,
+      "loss": 6.254307746887207,
+      "step": 4864
+    },
+    {
+      "epoch": 0.236478684587317,
+      "grad_norm": 3.2061522006988525,
+      "learning_rate": 0.0009874418626126245,
+      "loss": 6.149478912353516,
+      "step": 5120
+    },
+    {
+      "epoch": 0.236478684587317,
+      "eval_bleu": 0.5571576551016023,
+      "eval_ce_loss": 0.00175463656717316,
+      "eval_cos_loss": 0.31512473918259415,
+      "eval_cov_loss": 0.00032193904485315356,
+      "eval_flow_ce_loss": 1.9341952408829781,
+      "eval_loss": 6.142604702139554,
+      "eval_mean_loss": 0.0007544319254620485,
+      "eval_mse_loss": 1.029675681433177,
+      "eval_whiten_loss": 0.1808091167989931,
+      "flow/cos_sim": 0.684875273677312,
+      "flow/improvement_ratio": 0.9711803045447014,
+      "flow/mag_ratio_mean": 0.6440029551177264,
+      "flow/mag_ratio_std": 0.19927926647336516,
+      "step": 5120
+    },
+    {
+      "epoch": 0.236478684587317,
+      "eval_bleu": 0.5571576551016023,
+      "eval_ce_loss": 0.00175463656717316,
+      "eval_cos_loss": 0.31512473918259415,
+      "eval_cov_loss": 0.00032193904485315356,
+      "eval_flow_ce_loss": 1.9341952408829781,
+      "eval_loss": 6.142604702139554,
+      "eval_mean_loss": 0.0007544319254620485,
+      "eval_mse_loss": 1.029675681433177,
+      "eval_runtime": 135.3862,
+      "eval_samples_per_second": 206.764,
+      "eval_steps_per_second": 3.235,
+      "eval_whiten_loss": 0.1808091167989931,
+      "flow/cos_sim": 0.684875273677312,
+      "flow/improvement_ratio": 0.9711803045447014,
+      "flow/mag_ratio_mean": 0.6440029551177264,
+      "flow/mag_ratio_std": 0.19927926647336516,
+      "step": 5120
+    },
+    {
+      "epoch": 0.24830261881668284,
+      "grad_norm": 2.7209255695343018,
+      "learning_rate": 0.0009860141576279757,
+      "loss": 6.127026081085205,
+      "step": 5376
+    },
+    {
+      "epoch": 0.2601265530460487,
+      "grad_norm": 2.651538133621216,
+      "learning_rate": 0.000984510752201038,
+      "loss": 6.14462947845459,
+      "step": 5632
+    },
+    {
+      "epoch": 0.27195048727541454,
+      "grad_norm": 2.6247918605804443,
+      "learning_rate": 0.0009829318804987673,
+      "loss": 6.107431411743164,
+      "step": 5888
+    },
+    {
+      "epoch": 0.2837744215047804,
+      "grad_norm": 2.3757526874542236,
+      "learning_rate": 0.0009812777884425719,
+      "loss": 6.154008865356445,
+      "step": 6144
+    },
+    {
+      "epoch": 0.2837744215047804,
+      "eval_bleu": 0.5627678380598656,
+      "eval_ce_loss": 0.0014910807216579316,
+      "eval_cos_loss": 0.3033045921559748,
+      "eval_cov_loss": 0.00024372758732531347,
+      "eval_flow_ce_loss": 1.9364468174982288,
+      "eval_loss": 6.006853520598042,
+      "eval_mean_loss": 0.0007080443127500212,
+      "eval_mse_loss": 0.985551782802904,
+      "eval_whiten_loss": 0.13030236823373734,
+      "flow/cos_sim": 0.6966954118584934,
+      "flow/improvement_ratio": 0.9679767886797587,
+      "flow/mag_ratio_mean": 0.6562947229435455,
+      "flow/mag_ratio_std": 0.20180939535980355,
+      "step": 6144
+    },
+    {
+      "epoch": 0.2837744215047804,
+      "eval_bleu": 0.5627678380598656,
+      "eval_ce_loss": 0.0014910807216579316,
+      "eval_cos_loss": 0.3033045921559748,
+      "eval_cov_loss": 0.00024372758732531347,
+      "eval_flow_ce_loss": 1.9364468174982288,
+      "eval_loss": 6.006853520598042,
+      "eval_mean_loss": 0.0007080443127500212,
+      "eval_mse_loss": 0.985551782802904,
+      "eval_runtime": 142.3565,
+      "eval_samples_per_second": 196.64,
+      "eval_steps_per_second": 3.077,
+      "eval_whiten_loss": 0.13030236823373734,
+      "flow/cos_sim": 0.6966954118584934,
+      "flow/improvement_ratio": 0.9679767886797587,
+      "flow/mag_ratio_mean": 0.6562947229435455,
+      "flow/mag_ratio_std": 0.20180939535980355,
+      "step": 6144
+    },
+    {
+      "epoch": 0.2955983557341462,
+      "grad_norm": 3.0790116786956787,
+      "learning_rate": 0.0009795487336700087,
+      "loss": 6.061086177825928,
+      "step": 6400
+    },
+    {
+      "epoch": 0.30742228996351206,
+      "grad_norm": 3.054367780685425,
+      "learning_rate": 0.0009777449854946542,
+      "loss": 6.01887321472168,
+      "step": 6656
+    },
+    {
+      "epoch": 0.3192462241928779,
+      "grad_norm": 2.9489176273345947,
+      "learning_rate": 0.0009758668248641569,
+      "loss": 6.025762557983398,
+      "step": 6912
+    },
+    {
+      "epoch": 0.3310701584222438,
+      "grad_norm": 3.1671557426452637,
+      "learning_rate": 0.0009739145443164773,
+      "loss": 6.037273406982422,
+      "step": 7168
+    },
+    {
+      "epoch": 0.3310701584222438,
+      "eval_bleu": 0.5363273200410009,
+      "eval_ce_loss": 0.0013653414107839999,
+      "eval_cos_loss": 0.293757545927616,
+      "eval_cov_loss": 0.00026882515744317844,
+      "eval_flow_ce_loss": 1.9674101572602851,
+      "eval_loss": 6.0363281105206985,
+      "eval_mean_loss": 0.000960654606612093,
+      "eval_mse_loss": 0.9632442325489706,
+      "eval_whiten_loss": 0.14329069825612248,
+      "flow/cos_sim": 0.7062424644487634,
+      "flow/improvement_ratio": 0.9698878646441246,
+      "flow/mag_ratio_mean": 0.6735025889252964,
+      "flow/mag_ratio_std": 0.2078927657715806,
+      "step": 7168
+    },
+    {
+      "epoch": 0.3310701584222438,
+      "eval_bleu": 0.5363273200410009,
+      "eval_ce_loss": 0.0013653414107839999,
+      "eval_cos_loss": 0.293757545927616,
+      "eval_cov_loss": 0.00026882515744317844,
+      "eval_flow_ce_loss": 1.9674101572602851,
+      "eval_loss": 6.0363281105206985,
+      "eval_mean_loss": 0.000960654606612093,
+      "eval_mse_loss": 0.9632442325489706,
+      "eval_runtime": 144.5396,
+      "eval_samples_per_second": 193.67,
+      "eval_steps_per_second": 3.03,
+      "eval_whiten_loss": 0.14329069825612248,
+      "flow/cos_sim": 0.7062424644487634,
+      "flow/improvement_ratio": 0.9698878646441246,
+      "flow/mag_ratio_mean": 0.6735025889252964,
+      "flow/mag_ratio_std": 0.2078927657715806,
+      "step": 7168
+    }
+  ],
+  "logging_steps": 256,
+  "max_steps": 64953,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "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-v4.4+/checkpoint-7168/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6bce1f7f6d474249ed7d017dc2251184e884f8d0ae83eb277b6a88815e6d685a
+size 5137