Upload folder using huggingface_hub

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0f74d95c410dd375e45fb75e89e88f171bf3cb82d896571f983250787ef4376a
+size 379087640

optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:107fda308f0eee1d060e9b6a20682651e0976402cdf649cc472d32273816c008
+size 3152668

rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60c188633c8b74e994b9ed331d0da678d7e187f6befa9f01433e98f836aa17aa
+size 14308

scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7f3448c0ab334c9b7b6e99b96a0fa2bf6c7990574f81429f1a526818155658c
+size 1064

trainer_state.json

@@ -0,0 +1,2273 @@
+{
+  "best_metric": 0.13140406753766004,
+  "best_model_checkpoint": "/workspace/disk2/krishna/checkpoints/checkpoint-1390",
+  "epoch": 0.14,
+  "eval_steps": 10,
+  "global_step": 1400,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.001,
+      "grad_norm": 1.7856025695800781,
+      "learning_rate": 5.000000000000001e-07,
+      "loss": 0.682,
+      "step": 10
+    },
+    {
+      "epoch": 0.001,
+      "eval_cos_sim": 0.30435964465141296,
+      "eval_loss": 0.6965229106162732,
+      "eval_runtime": 181.358,
+      "eval_samples_per_second": 22.056,
+      "eval_steps_per_second": 0.347,
+      "step": 10
+    },
+    {
+      "epoch": 0.002,
+      "grad_norm": 1.632306694984436,
+      "learning_rate": 1.0000000000000002e-06,
+      "loss": 0.6902,
+      "step": 20
+    },
+    {
+      "epoch": 0.002,
+      "eval_cos_sim": 0.3060647249221802,
+      "eval_loss": 0.6948180051063245,
+      "eval_runtime": 158.9409,
+      "eval_samples_per_second": 25.167,
+      "eval_steps_per_second": 0.396,
+      "step": 20
+    },
+    {
+      "epoch": 0.003,
+      "grad_norm": 1.4781064987182617,
+      "learning_rate": 1.5e-06,
+      "loss": 0.6879,
+      "step": 30
+    },
+    {
+      "epoch": 0.003,
+      "eval_cos_sim": 0.30893880128860474,
+      "eval_loss": 0.6919441828941053,
+      "eval_runtime": 163.5656,
+      "eval_samples_per_second": 24.455,
+      "eval_steps_per_second": 0.385,
+      "step": 30
+    },
+    {
+      "epoch": 0.004,
+      "grad_norm": 1.4996154308319092,
+      "learning_rate": 2.0000000000000003e-06,
+      "loss": 0.6805,
+      "step": 40
+    },
+    {
+      "epoch": 0.004,
+      "eval_cos_sim": 0.3130052089691162,
+      "eval_loss": 0.6878782258247084,
+      "eval_runtime": 160.8804,
+      "eval_samples_per_second": 24.863,
+      "eval_steps_per_second": 0.392,
+      "step": 40
+    },
+    {
+      "epoch": 0.005,
+      "grad_norm": 1.7979084253311157,
+      "learning_rate": 2.5e-06,
+      "loss": 0.6795,
+      "step": 50
+    },
+    {
+      "epoch": 0.005,
+      "eval_cos_sim": 0.3181459903717041,
+      "eval_loss": 0.6827379918311781,
+      "eval_runtime": 159.4021,
+      "eval_samples_per_second": 25.094,
+      "eval_steps_per_second": 0.395,
+      "step": 50
+    },
+    {
+      "epoch": 0.006,
+      "grad_norm": 1.4049359560012817,
+      "learning_rate": 3e-06,
+      "loss": 0.6781,
+      "step": 60
+    },
+    {
+      "epoch": 0.006,
+      "eval_cos_sim": 0.32444360852241516,
+      "eval_loss": 0.6764405532096571,
+      "eval_runtime": 159.9515,
+      "eval_samples_per_second": 25.008,
+      "eval_steps_per_second": 0.394,
+      "step": 60
+    },
+    {
+      "epoch": 0.007,
+      "grad_norm": 1.7637789249420166,
+      "learning_rate": 3.5000000000000004e-06,
+      "loss": 0.6718,
+      "step": 70
+    },
+    {
+      "epoch": 0.007,
+      "eval_cos_sim": 0.331728994846344,
+      "eval_loss": 0.6691558842872327,
+      "eval_runtime": 160.8921,
+      "eval_samples_per_second": 24.861,
+      "eval_steps_per_second": 0.392,
+      "step": 70
+    },
+    {
+      "epoch": 0.008,
+      "grad_norm": 1.7560211420059204,
+      "learning_rate": 4.000000000000001e-06,
+      "loss": 0.6551,
+      "step": 80
+    },
+    {
+      "epoch": 0.008,
+      "eval_cos_sim": 0.3400111794471741,
+      "eval_loss": 0.6608743510459608,
+      "eval_runtime": 172.4109,
+      "eval_samples_per_second": 23.2,
+      "eval_steps_per_second": 0.365,
+      "step": 80
+    },
+    {
+      "epoch": 0.009,
+      "grad_norm": 1.7577630281448364,
+      "learning_rate": 4.5e-06,
+      "loss": 0.6566,
+      "step": 90
+    },
+    {
+      "epoch": 0.009,
+      "eval_cos_sim": 0.34923624992370605,
+      "eval_loss": 0.6516506343101209,
+      "eval_runtime": 161.459,
+      "eval_samples_per_second": 24.774,
+      "eval_steps_per_second": 0.39,
+      "step": 90
+    },
+    {
+      "epoch": 0.01,
+      "grad_norm": 1.7723366022109985,
+      "learning_rate": 5e-06,
+      "loss": 0.6332,
+      "step": 100
+    },
+    {
+      "epoch": 0.01,
+      "eval_cos_sim": 0.35936808586120605,
+      "eval_loss": 0.6415200128768629,
+      "eval_runtime": 158.9596,
+      "eval_samples_per_second": 25.164,
+      "eval_steps_per_second": 0.396,
+      "step": 100
+    },
+    {
+      "epoch": 0.011,
+      "grad_norm": 1.7818301916122437,
+      "learning_rate": 5.500000000000001e-06,
+      "loss": 0.6297,
+      "step": 110
+    },
+    {
+      "epoch": 0.011,
+      "eval_cos_sim": 0.3702617287635803,
+      "eval_loss": 0.6306272282813734,
+      "eval_runtime": 161.4739,
+      "eval_samples_per_second": 24.772,
+      "eval_steps_per_second": 0.39,
+      "step": 110
+    },
+    {
+      "epoch": 0.012,
+      "grad_norm": 1.4825223684310913,
+      "learning_rate": 6e-06,
+      "loss": 0.6473,
+      "step": 120
+    },
+    {
+      "epoch": 0.012,
+      "eval_cos_sim": 0.381907194852829,
+      "eval_loss": 0.6189834599708265,
+      "eval_runtime": 162.0142,
+      "eval_samples_per_second": 24.689,
+      "eval_steps_per_second": 0.389,
+      "step": 120
+    },
+    {
+      "epoch": 0.013,
+      "grad_norm": 1.7424395084381104,
+      "learning_rate": 6.5000000000000004e-06,
+      "loss": 0.611,
+      "step": 130
+    },
+    {
+      "epoch": 0.013,
+      "eval_cos_sim": 0.3940889537334442,
+      "eval_loss": 0.6068024029945082,
+      "eval_runtime": 161.0467,
+      "eval_samples_per_second": 24.838,
+      "eval_steps_per_second": 0.391,
+      "step": 130
+    },
+    {
+      "epoch": 0.014,
+      "grad_norm": 1.7474108934402466,
+      "learning_rate": 7.000000000000001e-06,
+      "loss": 0.5932,
+      "step": 140
+    },
+    {
+      "epoch": 0.014,
+      "eval_cos_sim": 0.4073576033115387,
+      "eval_loss": 0.5935340962623304,
+      "eval_runtime": 161.0017,
+      "eval_samples_per_second": 24.844,
+      "eval_steps_per_second": 0.391,
+      "step": 140
+    },
+    {
+      "epoch": 0.015,
+      "grad_norm": 1.6058062314987183,
+      "learning_rate": 7.5e-06,
+      "loss": 0.5824,
+      "step": 150
+    },
+    {
+      "epoch": 0.015,
+      "eval_cos_sim": 0.42149588465690613,
+      "eval_loss": 0.5793958764289564,
+      "eval_runtime": 167.7429,
+      "eval_samples_per_second": 23.846,
+      "eval_steps_per_second": 0.376,
+      "step": 150
+    },
+    {
+      "epoch": 0.016,
+      "grad_norm": 1.5788137912750244,
+      "learning_rate": 8.000000000000001e-06,
+      "loss": 0.5634,
+      "step": 160
+    },
+    {
+      "epoch": 0.016,
+      "eval_cos_sim": 0.43618762493133545,
+      "eval_loss": 0.564703805944795,
+      "eval_runtime": 159.9946,
+      "eval_samples_per_second": 25.001,
+      "eval_steps_per_second": 0.394,
+      "step": 160
+    },
+    {
+      "epoch": 0.017,
+      "grad_norm": 1.6113280057907104,
+      "learning_rate": 8.500000000000002e-06,
+      "loss": 0.5519,
+      "step": 170
+    },
+    {
+      "epoch": 0.017,
+      "eval_cos_sim": 0.4508235454559326,
+      "eval_loss": 0.5500685830329604,
+      "eval_runtime": 157.8438,
+      "eval_samples_per_second": 25.342,
+      "eval_steps_per_second": 0.399,
+      "step": 170
+    },
+    {
+      "epoch": 0.018,
+      "grad_norm": 1.6156154870986938,
+      "learning_rate": 9e-06,
+      "loss": 0.5407,
+      "step": 180
+    },
+    {
+      "epoch": 0.018,
+      "eval_cos_sim": 0.46566617488861084,
+      "eval_loss": 0.5352267394279188,
+      "eval_runtime": 159.9444,
+      "eval_samples_per_second": 25.009,
+      "eval_steps_per_second": 0.394,
+      "step": 180
+    },
+    {
+      "epoch": 0.019,
+      "grad_norm": 1.5958607196807861,
+      "learning_rate": 9.5e-06,
+      "loss": 0.5117,
+      "step": 190
+    },
+    {
+      "epoch": 0.019,
+      "eval_cos_sim": 0.4806951582431793,
+      "eval_loss": 0.5202000065063185,
+      "eval_runtime": 157.8685,
+      "eval_samples_per_second": 25.338,
+      "eval_steps_per_second": 0.399,
+      "step": 190
+    },
+    {
+      "epoch": 0.02,
+      "grad_norm": 1.6012651920318604,
+      "learning_rate": 1e-05,
+      "loss": 0.5116,
+      "step": 200
+    },
+    {
+      "epoch": 0.02,
+      "eval_cos_sim": 0.49558621644973755,
+      "eval_loss": 0.5053114848350233,
+      "eval_runtime": 155.395,
+      "eval_samples_per_second": 25.741,
+      "eval_steps_per_second": 0.405,
+      "step": 200
+    },
+    {
+      "epoch": 0.021,
+      "grad_norm": 1.7965919971466064,
+      "learning_rate": 1.05e-05,
+      "loss": 0.4782,
+      "step": 210
+    },
+    {
+      "epoch": 0.021,
+      "eval_cos_sim": 0.5108910202980042,
+      "eval_loss": 0.49000780775296154,
+      "eval_runtime": 159.6303,
+      "eval_samples_per_second": 25.058,
+      "eval_steps_per_second": 0.395,
+      "step": 210
+    },
+    {
+      "epoch": 0.022,
+      "grad_norm": 1.791284441947937,
+      "learning_rate": 1.1000000000000001e-05,
+      "loss": 0.4802,
+      "step": 220
+    },
+    {
+      "epoch": 0.022,
+      "eval_cos_sim": 0.5271998643875122,
+      "eval_loss": 0.4737012028907484,
+      "eval_runtime": 169.963,
+      "eval_samples_per_second": 23.535,
+      "eval_steps_per_second": 0.371,
+      "step": 220
+    },
+    {
+      "epoch": 0.023,
+      "grad_norm": 1.7923716306686401,
+      "learning_rate": 1.1500000000000002e-05,
+      "loss": 0.4513,
+      "step": 230
+    },
+    {
+      "epoch": 0.023,
+      "eval_cos_sim": 0.5435076355934143,
+      "eval_loss": 0.45739728691327036,
+      "eval_runtime": 160.753,
+      "eval_samples_per_second": 24.883,
+      "eval_steps_per_second": 0.392,
+      "step": 230
+    },
+    {
+      "epoch": 0.024,
+      "grad_norm": 1.7997322082519531,
+      "learning_rate": 1.2e-05,
+      "loss": 0.4288,
+      "step": 240
+    },
+    {
+      "epoch": 0.024,
+      "eval_cos_sim": 0.5595658421516418,
+      "eval_loss": 0.44134216930615366,
+      "eval_runtime": 163.2069,
+      "eval_samples_per_second": 24.509,
+      "eval_steps_per_second": 0.386,
+      "step": 240
+    },
+    {
+      "epoch": 0.025,
+      "grad_norm": 1.7595293521881104,
+      "learning_rate": 1.25e-05,
+      "loss": 0.4365,
+      "step": 250
+    },
+    {
+      "epoch": 0.025,
+      "eval_cos_sim": 0.5746485590934753,
+      "eval_loss": 0.4262613997672743,
+      "eval_runtime": 160.8892,
+      "eval_samples_per_second": 24.862,
+      "eval_steps_per_second": 0.392,
+      "step": 250
+    },
+    {
+      "epoch": 0.026,
+      "grad_norm": 1.614985704421997,
+      "learning_rate": 1.3000000000000001e-05,
+      "loss": 0.4007,
+      "step": 260
+    },
+    {
+      "epoch": 0.026,
+      "eval_cos_sim": 0.5887113809585571,
+      "eval_loss": 0.4121987219070143,
+      "eval_runtime": 157.6029,
+      "eval_samples_per_second": 25.38,
+      "eval_steps_per_second": 0.4,
+      "step": 260
+    },
+    {
+      "epoch": 0.027,
+      "grad_norm": 1.6005358695983887,
+      "learning_rate": 1.3500000000000001e-05,
+      "loss": 0.4048,
+      "step": 270
+    },
+    {
+      "epoch": 0.027,
+      "eval_cos_sim": 0.6018706560134888,
+      "eval_loss": 0.39904179241406385,
+      "eval_runtime": 163.2521,
+      "eval_samples_per_second": 24.502,
+      "eval_steps_per_second": 0.386,
+      "step": 270
+    },
+    {
+      "epoch": 0.028,
+      "grad_norm": 1.7216198444366455,
+      "learning_rate": 1.4000000000000001e-05,
+      "loss": 0.3793,
+      "step": 280
+    },
+    {
+      "epoch": 0.028,
+      "eval_cos_sim": 0.6152481436729431,
+      "eval_loss": 0.385668321153993,
+      "eval_runtime": 163.9341,
+      "eval_samples_per_second": 24.4,
+      "eval_steps_per_second": 0.384,
+      "step": 280
+    },
+    {
+      "epoch": 0.029,
+      "grad_norm": 1.7449027299880981,
+      "learning_rate": 1.45e-05,
+      "loss": 0.3614,
+      "step": 290
+    },
+    {
+      "epoch": 0.029,
+      "eval_cos_sim": 0.6277655959129333,
+      "eval_loss": 0.3731525359367079,
+      "eval_runtime": 170.8176,
+      "eval_samples_per_second": 23.417,
+      "eval_steps_per_second": 0.369,
+      "step": 290
+    },
+    {
+      "epoch": 0.03,
+      "grad_norm": 1.5057954788208008,
+      "learning_rate": 1.5e-05,
+      "loss": 0.3557,
+      "step": 300
+    },
+    {
+      "epoch": 0.03,
+      "eval_cos_sim": 0.6390805244445801,
+      "eval_loss": 0.36183924534069956,
+      "eval_runtime": 160.3353,
+      "eval_samples_per_second": 24.948,
+      "eval_steps_per_second": 0.393,
+      "step": 300
+    },
+    {
+      "epoch": 0.031,
+      "grad_norm": 1.5245262384414673,
+      "learning_rate": 1.55e-05,
+      "loss": 0.3729,
+      "step": 310
+    },
+    {
+      "epoch": 0.031,
+      "eval_cos_sim": 0.6503457427024841,
+      "eval_loss": 0.3505764603828138,
+      "eval_runtime": 161.447,
+      "eval_samples_per_second": 24.776,
+      "eval_steps_per_second": 0.39,
+      "step": 310
+    },
+    {
+      "epoch": 0.032,
+      "grad_norm": 1.4508600234985352,
+      "learning_rate": 1.6000000000000003e-05,
+      "loss": 0.3476,
+      "step": 320
+    },
+    {
+      "epoch": 0.032,
+      "eval_cos_sim": 0.664078414440155,
+      "eval_loss": 0.33684703972088753,
+      "eval_runtime": 162.7312,
+      "eval_samples_per_second": 24.58,
+      "eval_steps_per_second": 0.387,
+      "step": 320
+    },
+    {
+      "epoch": 0.033,
+      "grad_norm": 1.545847773551941,
+      "learning_rate": 1.65e-05,
+      "loss": 0.3335,
+      "step": 330
+    },
+    {
+      "epoch": 0.033,
+      "eval_cos_sim": 0.6734265089035034,
+      "eval_loss": 0.32750138094174325,
+      "eval_runtime": 162.4883,
+      "eval_samples_per_second": 24.617,
+      "eval_steps_per_second": 0.388,
+      "step": 330
+    },
+    {
+      "epoch": 0.034,
+      "grad_norm": 1.5199618339538574,
+      "learning_rate": 1.7000000000000003e-05,
+      "loss": 0.3369,
+      "step": 340
+    },
+    {
+      "epoch": 0.034,
+      "eval_cos_sim": 0.6810038685798645,
+      "eval_loss": 0.31992570402371345,
+      "eval_runtime": 161.4031,
+      "eval_samples_per_second": 24.783,
+      "eval_steps_per_second": 0.39,
+      "step": 340
+    },
+    {
+      "epoch": 0.035,
+      "grad_norm": 1.2642650604248047,
+      "learning_rate": 1.75e-05,
+      "loss": 0.3396,
+      "step": 350
+    },
+    {
+      "epoch": 0.035,
+      "eval_cos_sim": 0.6911267638206482,
+      "eval_loss": 0.3098063874458021,
+      "eval_runtime": 165.9265,
+      "eval_samples_per_second": 24.107,
+      "eval_steps_per_second": 0.38,
+      "step": 350
+    },
+    {
+      "epoch": 0.036,
+      "grad_norm": 1.2891690731048584,
+      "learning_rate": 1.8e-05,
+      "loss": 0.3015,
+      "step": 360
+    },
+    {
+      "epoch": 0.036,
+      "eval_cos_sim": 0.6987318992614746,
+      "eval_loss": 0.3022033410285658,
+      "eval_runtime": 173.6837,
+      "eval_samples_per_second": 23.03,
+      "eval_steps_per_second": 0.363,
+      "step": 360
+    },
+    {
+      "epoch": 0.037,
+      "grad_norm": 1.2396866083145142,
+      "learning_rate": 1.85e-05,
+      "loss": 0.3077,
+      "step": 370
+    },
+    {
+      "epoch": 0.037,
+      "eval_cos_sim": 0.7082959413528442,
+      "eval_loss": 0.2926409912322706,
+      "eval_runtime": 158.0401,
+      "eval_samples_per_second": 25.31,
+      "eval_steps_per_second": 0.399,
+      "step": 370
+    },
+    {
+      "epoch": 0.038,
+      "grad_norm": 1.2055168151855469,
+      "learning_rate": 1.9e-05,
+      "loss": 0.2895,
+      "step": 380
+    },
+    {
+      "epoch": 0.038,
+      "eval_cos_sim": 0.7168179750442505,
+      "eval_loss": 0.284124677202577,
+      "eval_runtime": 160.1742,
+      "eval_samples_per_second": 24.973,
+      "eval_steps_per_second": 0.393,
+      "step": 380
+    },
+    {
+      "epoch": 0.039,
+      "grad_norm": 1.1191834211349487,
+      "learning_rate": 1.9500000000000003e-05,
+      "loss": 0.2944,
+      "step": 390
+    },
+    {
+      "epoch": 0.039,
+      "eval_cos_sim": 0.7234470248222351,
+      "eval_loss": 0.2775011363242811,
+      "eval_runtime": 165.4196,
+      "eval_samples_per_second": 24.181,
+      "eval_steps_per_second": 0.381,
+      "step": 390
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 1.097105860710144,
+      "learning_rate": 2e-05,
+      "loss": 0.2898,
+      "step": 400
+    },
+    {
+      "epoch": 0.04,
+      "eval_cos_sim": 0.7314009666442871,
+      "eval_loss": 0.2695479884360975,
+      "eval_runtime": 159.5742,
+      "eval_samples_per_second": 25.067,
+      "eval_steps_per_second": 0.395,
+      "step": 400
+    },
+    {
+      "epoch": 0.041,
+      "grad_norm": 1.040711760520935,
+      "learning_rate": 2.05e-05,
+      "loss": 0.2573,
+      "step": 410
+    },
+    {
+      "epoch": 0.041,
+      "eval_cos_sim": 0.7389957904815674,
+      "eval_loss": 0.26195128538357676,
+      "eval_runtime": 160.3231,
+      "eval_samples_per_second": 24.95,
+      "eval_steps_per_second": 0.393,
+      "step": 410
+    },
+    {
+      "epoch": 0.042,
+      "grad_norm": 1.0322086811065674,
+      "learning_rate": 2.1e-05,
+      "loss": 0.241,
+      "step": 420
+    },
+    {
+      "epoch": 0.042,
+      "eval_cos_sim": 0.7471259236335754,
+      "eval_loss": 0.25382448627697884,
+      "eval_runtime": 161.5365,
+      "eval_samples_per_second": 24.762,
+      "eval_steps_per_second": 0.39,
+      "step": 420
+    },
+    {
+      "epoch": 0.043,
+      "grad_norm": 0.9718736410140991,
+      "learning_rate": 2.15e-05,
+      "loss": 0.26,
+      "step": 430
+    },
+    {
+      "epoch": 0.043,
+      "eval_cos_sim": 0.7534591555595398,
+      "eval_loss": 0.24748690702664317,
+      "eval_runtime": 169.2487,
+      "eval_samples_per_second": 23.634,
+      "eval_steps_per_second": 0.372,
+      "step": 430
+    },
+    {
+      "epoch": 0.044,
+      "grad_norm": 0.8780993223190308,
+      "learning_rate": 2.2000000000000003e-05,
+      "loss": 0.2255,
+      "step": 440
+    },
+    {
+      "epoch": 0.044,
+      "eval_cos_sim": 0.7604101896286011,
+      "eval_loss": 0.2405370173667616,
+      "eval_runtime": 161.7354,
+      "eval_samples_per_second": 24.732,
+      "eval_steps_per_second": 0.39,
+      "step": 440
+    },
+    {
+      "epoch": 0.045,
+      "grad_norm": 0.8193601369857788,
+      "learning_rate": 2.25e-05,
+      "loss": 0.2182,
+      "step": 450
+    },
+    {
+      "epoch": 0.045,
+      "eval_cos_sim": 0.7673761248588562,
+      "eval_loss": 0.2335676951621718,
+      "eval_runtime": 159.2952,
+      "eval_samples_per_second": 25.111,
+      "eval_steps_per_second": 0.395,
+      "step": 450
+    },
+    {
+      "epoch": 0.046,
+      "grad_norm": 0.8820355534553528,
+      "learning_rate": 2.3000000000000003e-05,
+      "loss": 0.2153,
+      "step": 460
+    },
+    {
+      "epoch": 0.046,
+      "eval_cos_sim": 0.7723870277404785,
+      "eval_loss": 0.228554349443788,
+      "eval_runtime": 161.8756,
+      "eval_samples_per_second": 24.71,
+      "eval_steps_per_second": 0.389,
+      "step": 460
+    },
+    {
+      "epoch": 0.047,
+      "grad_norm": 0.9045171141624451,
+      "learning_rate": 2.35e-05,
+      "loss": 0.2245,
+      "step": 470
+    },
+    {
+      "epoch": 0.047,
+      "eval_cos_sim": 0.7763780951499939,
+      "eval_loss": 0.22456564476239152,
+      "eval_runtime": 160.6855,
+      "eval_samples_per_second": 24.893,
+      "eval_steps_per_second": 0.392,
+      "step": 470
+    },
+    {
+      "epoch": 0.048,
+      "grad_norm": 0.7304402589797974,
+      "learning_rate": 2.4e-05,
+      "loss": 0.2069,
+      "step": 480
+    },
+    {
+      "epoch": 0.048,
+      "eval_cos_sim": 0.7830195426940918,
+      "eval_loss": 0.21792557956921524,
+      "eval_runtime": 162.6536,
+      "eval_samples_per_second": 24.592,
+      "eval_steps_per_second": 0.387,
+      "step": 480
+    },
+    {
+      "epoch": 0.049,
+      "grad_norm": 0.6550305485725403,
+      "learning_rate": 2.45e-05,
+      "loss": 0.2001,
+      "step": 490
+    },
+    {
+      "epoch": 0.049,
+      "eval_cos_sim": 0.7868452072143555,
+      "eval_loss": 0.21410264875637955,
+      "eval_runtime": 163.12,
+      "eval_samples_per_second": 24.522,
+      "eval_steps_per_second": 0.386,
+      "step": 490
+    },
+    {
+      "epoch": 0.05,
+      "grad_norm": 0.6369508504867554,
+      "learning_rate": 2.5e-05,
+      "loss": 0.1935,
+      "step": 500
+    },
+    {
+      "epoch": 0.05,
+      "eval_cos_sim": 0.7914958596229553,
+      "eval_loss": 0.2094592919563002,
+      "eval_runtime": 164.3494,
+      "eval_samples_per_second": 24.338,
+      "eval_steps_per_second": 0.383,
+      "step": 500
+    },
+    {
+      "epoch": 0.051,
+      "grad_norm": 0.6336702108383179,
+      "learning_rate": 2.5500000000000003e-05,
+      "loss": 0.2018,
+      "step": 510
+    },
+    {
+      "epoch": 0.051,
+      "eval_cos_sim": 0.7958762645721436,
+      "eval_loss": 0.20508341314541764,
+      "eval_runtime": 163.0667,
+      "eval_samples_per_second": 24.53,
+      "eval_steps_per_second": 0.386,
+      "step": 510
+    },
+    {
+      "epoch": 0.052,
+      "grad_norm": 0.585249125957489,
+      "learning_rate": 2.6000000000000002e-05,
+      "loss": 0.1854,
+      "step": 520
+    },
+    {
+      "epoch": 0.052,
+      "eval_cos_sim": 0.7985916137695312,
+      "eval_loss": 0.20236765434491105,
+      "eval_runtime": 161.7682,
+      "eval_samples_per_second": 24.727,
+      "eval_steps_per_second": 0.389,
+      "step": 520
+    },
+    {
+      "epoch": 0.053,
+      "grad_norm": 0.5437793731689453,
+      "learning_rate": 2.6500000000000004e-05,
+      "loss": 0.1795,
+      "step": 530
+    },
+    {
+      "epoch": 0.053,
+      "eval_cos_sim": 0.8020971417427063,
+      "eval_loss": 0.19886508896099991,
+      "eval_runtime": 164.1233,
+      "eval_samples_per_second": 24.372,
+      "eval_steps_per_second": 0.384,
+      "step": 530
+    },
+    {
+      "epoch": 0.054,
+      "grad_norm": 0.6123674511909485,
+      "learning_rate": 2.7000000000000002e-05,
+      "loss": 0.1807,
+      "step": 540
+    },
+    {
+      "epoch": 0.054,
+      "eval_cos_sim": 0.8049777150154114,
+      "eval_loss": 0.19599416496502822,
+      "eval_runtime": 163.7538,
+      "eval_samples_per_second": 24.427,
+      "eval_steps_per_second": 0.385,
+      "step": 540
+    },
+    {
+      "epoch": 0.055,
+      "grad_norm": 0.5047230124473572,
+      "learning_rate": 2.7500000000000004e-05,
+      "loss": 0.1769,
+      "step": 550
+    },
+    {
+      "epoch": 0.055,
+      "eval_cos_sim": 0.8080028295516968,
+      "eval_loss": 0.19296009018170304,
+      "eval_runtime": 158.2319,
+      "eval_samples_per_second": 25.279,
+      "eval_steps_per_second": 0.398,
+      "step": 550
+    },
+    {
+      "epoch": 0.056,
+      "grad_norm": 0.4361515939235687,
+      "learning_rate": 2.8000000000000003e-05,
+      "loss": 0.1783,
+      "step": 560
+    },
+    {
+      "epoch": 0.056,
+      "eval_cos_sim": 0.8107978701591492,
+      "eval_loss": 0.19017480518567031,
+      "eval_runtime": 164.6869,
+      "eval_samples_per_second": 24.289,
+      "eval_steps_per_second": 0.383,
+      "step": 560
+    },
+    {
+      "epoch": 0.057,
+      "grad_norm": 0.4613172113895416,
+      "learning_rate": 2.8499999999999998e-05,
+      "loss": 0.1714,
+      "step": 570
+    },
+    {
+      "epoch": 0.057,
+      "eval_cos_sim": 0.8138161301612854,
+      "eval_loss": 0.18716797020184464,
+      "eval_runtime": 167.1324,
+      "eval_samples_per_second": 23.933,
+      "eval_steps_per_second": 0.377,
+      "step": 570
+    },
+    {
+      "epoch": 0.058,
+      "grad_norm": 0.4287095069885254,
+      "learning_rate": 2.9e-05,
+      "loss": 0.1633,
+      "step": 580
+    },
+    {
+      "epoch": 0.058,
+      "eval_cos_sim": 0.8161399364471436,
+      "eval_loss": 0.1848432316993422,
+      "eval_runtime": 160.6945,
+      "eval_samples_per_second": 24.892,
+      "eval_steps_per_second": 0.392,
+      "step": 580
+    },
+    {
+      "epoch": 0.059,
+      "grad_norm": 0.442719429731369,
+      "learning_rate": 2.95e-05,
+      "loss": 0.1643,
+      "step": 590
+    },
+    {
+      "epoch": 0.059,
+      "eval_cos_sim": 0.8180400133132935,
+      "eval_loss": 0.1829425740455336,
+      "eval_runtime": 163.6004,
+      "eval_samples_per_second": 24.45,
+      "eval_steps_per_second": 0.385,
+      "step": 590
+    },
+    {
+      "epoch": 0.06,
+      "grad_norm": 0.38597649335861206,
+      "learning_rate": 3e-05,
+      "loss": 0.1645,
+      "step": 600
+    },
+    {
+      "epoch": 0.06,
+      "eval_cos_sim": 0.8204721808433533,
+      "eval_loss": 0.1805112529014296,
+      "eval_runtime": 158.8011,
+      "eval_samples_per_second": 25.189,
+      "eval_steps_per_second": 0.397,
+      "step": 600
+    },
+    {
+      "epoch": 0.061,
+      "grad_norm": 0.40187621116638184,
+      "learning_rate": 3.05e-05,
+      "loss": 0.1632,
+      "step": 610
+    },
+    {
+      "epoch": 0.061,
+      "eval_cos_sim": 0.8225390911102295,
+      "eval_loss": 0.17843791820752097,
+      "eval_runtime": 164.1328,
+      "eval_samples_per_second": 24.371,
+      "eval_steps_per_second": 0.384,
+      "step": 610
+    },
+    {
+      "epoch": 0.062,
+      "grad_norm": 0.36652931571006775,
+      "learning_rate": 3.1e-05,
+      "loss": 0.1587,
+      "step": 620
+    },
+    {
+      "epoch": 0.062,
+      "eval_cos_sim": 0.8244653344154358,
+      "eval_loss": 0.17650788643109275,
+      "eval_runtime": 160.7137,
+      "eval_samples_per_second": 24.889,
+      "eval_steps_per_second": 0.392,
+      "step": 620
+    },
+    {
+      "epoch": 0.063,
+      "grad_norm": 0.39730513095855713,
+      "learning_rate": 3.15e-05,
+      "loss": 0.164,
+      "step": 630
+    },
+    {
+      "epoch": 0.063,
+      "eval_cos_sim": 0.8262067437171936,
+      "eval_loss": 0.17477095749127342,
+      "eval_runtime": 165.8251,
+      "eval_samples_per_second": 24.122,
+      "eval_steps_per_second": 0.38,
+      "step": 630
+    },
+    {
+      "epoch": 0.064,
+      "grad_norm": 0.37115684151649475,
+      "learning_rate": 3.2000000000000005e-05,
+      "loss": 0.1562,
+      "step": 640
+    },
+    {
+      "epoch": 0.064,
+      "eval_cos_sim": 0.8285937905311584,
+      "eval_loss": 0.17237986900555563,
+      "eval_runtime": 166.1474,
+      "eval_samples_per_second": 24.075,
+      "eval_steps_per_second": 0.379,
+      "step": 640
+    },
+    {
+      "epoch": 0.065,
+      "grad_norm": 0.3438771367073059,
+      "learning_rate": 3.2500000000000004e-05,
+      "loss": 0.1628,
+      "step": 650
+    },
+    {
+      "epoch": 0.065,
+      "eval_cos_sim": 0.8305408358573914,
+      "eval_loss": 0.17043518688427878,
+      "eval_runtime": 156.9198,
+      "eval_samples_per_second": 25.491,
+      "eval_steps_per_second": 0.401,
+      "step": 650
+    },
+    {
+      "epoch": 0.066,
+      "grad_norm": 0.445515513420105,
+      "learning_rate": 3.3e-05,
+      "loss": 0.1588,
+      "step": 660
+    },
+    {
+      "epoch": 0.066,
+      "eval_cos_sim": 0.8313919305801392,
+      "eval_loss": 0.1695925650809951,
+      "eval_runtime": 160.5009,
+      "eval_samples_per_second": 24.922,
+      "eval_steps_per_second": 0.393,
+      "step": 660
+    },
+    {
+      "epoch": 0.067,
+      "grad_norm": 0.30833250284194946,
+      "learning_rate": 3.35e-05,
+      "loss": 0.1597,
+      "step": 670
+    },
+    {
+      "epoch": 0.067,
+      "eval_cos_sim": 0.8327498435974121,
+      "eval_loss": 0.16823711063610983,
+      "eval_runtime": 163.7947,
+      "eval_samples_per_second": 24.421,
+      "eval_steps_per_second": 0.385,
+      "step": 670
+    },
+    {
+      "epoch": 0.068,
+      "grad_norm": 0.30146872997283936,
+      "learning_rate": 3.4000000000000007e-05,
+      "loss": 0.1521,
+      "step": 680
+    },
+    {
+      "epoch": 0.068,
+      "eval_cos_sim": 0.8345705270767212,
+      "eval_loss": 0.16642027666317893,
+      "eval_runtime": 163.0767,
+      "eval_samples_per_second": 24.528,
+      "eval_steps_per_second": 0.386,
+      "step": 680
+    },
+    {
+      "epoch": 0.069,
+      "grad_norm": 0.3060350716114044,
+      "learning_rate": 3.45e-05,
+      "loss": 0.1499,
+      "step": 690
+    },
+    {
+      "epoch": 0.069,
+      "eval_cos_sim": 0.8357890844345093,
+      "eval_loss": 0.16520177223431537,
+      "eval_runtime": 161.8292,
+      "eval_samples_per_second": 24.717,
+      "eval_steps_per_second": 0.389,
+      "step": 690
+    },
+    {
+      "epoch": 0.07,
+      "grad_norm": 0.25746750831604004,
+      "learning_rate": 3.5e-05,
+      "loss": 0.1494,
+      "step": 700
+    },
+    {
+      "epoch": 0.07,
+      "eval_cos_sim": 0.8371787071228027,
+      "eval_loss": 0.16381443121182396,
+      "eval_runtime": 171.1282,
+      "eval_samples_per_second": 23.374,
+      "eval_steps_per_second": 0.368,
+      "step": 700
+    },
+    {
+      "epoch": 0.071,
+      "grad_norm": 0.27807655930519104,
+      "learning_rate": 3.55e-05,
+      "loss": 0.1503,
+      "step": 710
+    },
+    {
+      "epoch": 0.071,
+      "eval_cos_sim": 0.8385919332504272,
+      "eval_loss": 0.16239957573162986,
+      "eval_runtime": 173.6991,
+      "eval_samples_per_second": 23.028,
+      "eval_steps_per_second": 0.363,
+      "step": 710
+    },
+    {
+      "epoch": 0.072,
+      "grad_norm": 0.2822527587413788,
+      "learning_rate": 3.6e-05,
+      "loss": 0.1596,
+      "step": 720
+    },
+    {
+      "epoch": 0.072,
+      "eval_cos_sim": 0.8399993777275085,
+      "eval_loss": 0.16099353077160786,
+      "eval_runtime": 166.2107,
+      "eval_samples_per_second": 24.066,
+      "eval_steps_per_second": 0.379,
+      "step": 720
+    },
+    {
+      "epoch": 0.073,
+      "grad_norm": 0.3689976632595062,
+      "learning_rate": 3.65e-05,
+      "loss": 0.1522,
+      "step": 730
+    },
+    {
+      "epoch": 0.073,
+      "eval_cos_sim": 0.8407827615737915,
+      "eval_loss": 0.16021420147168108,
+      "eval_runtime": 163.1982,
+      "eval_samples_per_second": 24.51,
+      "eval_steps_per_second": 0.386,
+      "step": 730
+    },
+    {
+      "epoch": 0.074,
+      "grad_norm": 0.24420756101608276,
+      "learning_rate": 3.7e-05,
+      "loss": 0.1464,
+      "step": 740
+    },
+    {
+      "epoch": 0.074,
+      "eval_cos_sim": 0.8423337936401367,
+      "eval_loss": 0.15866206934201188,
+      "eval_runtime": 162.3325,
+      "eval_samples_per_second": 24.641,
+      "eval_steps_per_second": 0.388,
+      "step": 740
+    },
+    {
+      "epoch": 0.075,
+      "grad_norm": 0.2716599404811859,
+      "learning_rate": 3.7500000000000003e-05,
+      "loss": 0.1458,
+      "step": 750
+    },
+    {
+      "epoch": 0.075,
+      "eval_cos_sim": 0.8430612683296204,
+      "eval_loss": 0.157932794592256,
+      "eval_runtime": 163.6706,
+      "eval_samples_per_second": 24.439,
+      "eval_steps_per_second": 0.385,
+      "step": 750
+    },
+    {
+      "epoch": 0.076,
+      "grad_norm": 0.27678823471069336,
+      "learning_rate": 3.8e-05,
+      "loss": 0.1446,
+      "step": 760
+    },
+    {
+      "epoch": 0.076,
+      "eval_cos_sim": 0.8438640236854553,
+      "eval_loss": 0.15713081218945452,
+      "eval_runtime": 159.6959,
+      "eval_samples_per_second": 25.048,
+      "eval_steps_per_second": 0.394,
+      "step": 760
+    },
+    {
+      "epoch": 0.077,
+      "grad_norm": 0.23866839706897736,
+      "learning_rate": 3.85e-05,
+      "loss": 0.1487,
+      "step": 770
+    },
+    {
+      "epoch": 0.077,
+      "eval_cos_sim": 0.8451653718948364,
+      "eval_loss": 0.15582896663891738,
+      "eval_runtime": 170.448,
+      "eval_samples_per_second": 23.468,
+      "eval_steps_per_second": 0.37,
+      "step": 770
+    },
+    {
+      "epoch": 0.078,
+      "grad_norm": 0.24985139071941376,
+      "learning_rate": 3.9000000000000006e-05,
+      "loss": 0.143,
+      "step": 780
+    },
+    {
+      "epoch": 0.078,
+      "eval_cos_sim": 0.8460158705711365,
+      "eval_loss": 0.15497446682202284,
+      "eval_runtime": 168.3573,
+      "eval_samples_per_second": 23.759,
+      "eval_steps_per_second": 0.374,
+      "step": 780
+    },
+    {
+      "epoch": 0.079,
+      "grad_norm": 0.2435818761587143,
+      "learning_rate": 3.9500000000000005e-05,
+      "loss": 0.1464,
+      "step": 790
+    },
+    {
+      "epoch": 0.079,
+      "eval_cos_sim": 0.8470072150230408,
+      "eval_loss": 0.15398560383068982,
+      "eval_runtime": 163.7823,
+      "eval_samples_per_second": 24.423,
+      "eval_steps_per_second": 0.385,
+      "step": 790
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.25132063031196594,
+      "learning_rate": 4e-05,
+      "loss": 0.1463,
+      "step": 800
+    },
+    {
+      "epoch": 0.08,
+      "eval_cos_sim": 0.8480408191680908,
+      "eval_loss": 0.15294421245800913,
+      "eval_runtime": 160.63,
+      "eval_samples_per_second": 24.902,
+      "eval_steps_per_second": 0.392,
+      "step": 800
+    },
+    {
+      "epoch": 0.081,
+      "grad_norm": 0.22080835700035095,
+      "learning_rate": 4.05e-05,
+      "loss": 0.1449,
+      "step": 810
+    },
+    {
+      "epoch": 0.081,
+      "eval_cos_sim": 0.8484798669815063,
+      "eval_loss": 0.15251411964642467,
+      "eval_runtime": 165.2445,
+      "eval_samples_per_second": 24.207,
+      "eval_steps_per_second": 0.381,
+      "step": 810
+    },
+    {
+      "epoch": 0.082,
+      "grad_norm": 0.21103359758853912,
+      "learning_rate": 4.1e-05,
+      "loss": 0.1425,
+      "step": 820
+    },
+    {
+      "epoch": 0.082,
+      "eval_cos_sim": 0.8491047024726868,
+      "eval_loss": 0.15190097334134042,
+      "eval_runtime": 165.2932,
+      "eval_samples_per_second": 24.199,
+      "eval_steps_per_second": 0.381,
+      "step": 820
+    },
+    {
+      "epoch": 0.083,
+      "grad_norm": 0.4737427234649658,
+      "learning_rate": 4.15e-05,
+      "loss": 0.1451,
+      "step": 830
+    },
+    {
+      "epoch": 0.083,
+      "eval_cos_sim": 0.8501515984535217,
+      "eval_loss": 0.15085182239758432,
+      "eval_runtime": 164.4705,
+      "eval_samples_per_second": 24.32,
+      "eval_steps_per_second": 0.383,
+      "step": 830
+    },
+    {
+      "epoch": 0.084,
+      "grad_norm": 0.19636596739292145,
+      "learning_rate": 4.2e-05,
+      "loss": 0.139,
+      "step": 840
+    },
+    {
+      "epoch": 0.084,
+      "eval_cos_sim": 0.8513444066047668,
+      "eval_loss": 0.14965240146862924,
+      "eval_runtime": 168.3369,
+      "eval_samples_per_second": 23.762,
+      "eval_steps_per_second": 0.374,
+      "step": 840
+    },
+    {
+      "epoch": 0.085,
+      "grad_norm": 0.22018375992774963,
+      "learning_rate": 4.25e-05,
+      "loss": 0.139,
+      "step": 850
+    },
+    {
+      "epoch": 0.085,
+      "eval_cos_sim": 0.8519036769866943,
+      "eval_loss": 0.14909614183651865,
+      "eval_runtime": 170.0334,
+      "eval_samples_per_second": 23.525,
+      "eval_steps_per_second": 0.371,
+      "step": 850
+    },
+    {
+      "epoch": 0.086,
+      "grad_norm": 0.21147912740707397,
+      "learning_rate": 4.3e-05,
+      "loss": 0.1383,
+      "step": 860
+    },
+    {
+      "epoch": 0.086,
+      "eval_cos_sim": 0.8522551655769348,
+      "eval_loss": 0.14874456455456675,
+      "eval_runtime": 160.1434,
+      "eval_samples_per_second": 24.978,
+      "eval_steps_per_second": 0.393,
+      "step": 860
+    },
+    {
+      "epoch": 0.087,
+      "grad_norm": 0.20831769704818726,
+      "learning_rate": 4.35e-05,
+      "loss": 0.1386,
+      "step": 870
+    },
+    {
+      "epoch": 0.087,
+      "eval_cos_sim": 0.8532180786132812,
+      "eval_loss": 0.1477800307487196,
+      "eval_runtime": 170.3934,
+      "eval_samples_per_second": 23.475,
+      "eval_steps_per_second": 0.37,
+      "step": 870
+    },
+    {
+      "epoch": 0.088,
+      "grad_norm": 0.3051137328147888,
+      "learning_rate": 4.4000000000000006e-05,
+      "loss": 0.141,
+      "step": 880
+    },
+    {
+      "epoch": 0.088,
+      "eval_cos_sim": 0.8542919754981995,
+      "eval_loss": 0.14670453121411264,
+      "eval_runtime": 165.3769,
+      "eval_samples_per_second": 24.187,
+      "eval_steps_per_second": 0.381,
+      "step": 880
+    },
+    {
+      "epoch": 0.089,
+      "grad_norm": 0.19627386331558228,
+      "learning_rate": 4.4500000000000004e-05,
+      "loss": 0.1423,
+      "step": 890
+    },
+    {
+      "epoch": 0.089,
+      "eval_cos_sim": 0.8539001941680908,
+      "eval_loss": 0.1470875745032972,
+      "eval_runtime": 163.4789,
+      "eval_samples_per_second": 24.468,
+      "eval_steps_per_second": 0.385,
+      "step": 890
+    },
+    {
+      "epoch": 0.09,
+      "grad_norm": 0.1867413967847824,
+      "learning_rate": 4.5e-05,
+      "loss": 0.1371,
+      "step": 900
+    },
+    {
+      "epoch": 0.09,
+      "eval_cos_sim": 0.8544667959213257,
+      "eval_loss": 0.1465196500037855,
+      "eval_runtime": 159.8165,
+      "eval_samples_per_second": 25.029,
+      "eval_steps_per_second": 0.394,
+      "step": 900
+    },
+    {
+      "epoch": 0.091,
+      "grad_norm": 0.21700890362262726,
+      "learning_rate": 4.55e-05,
+      "loss": 0.1405,
+      "step": 910
+    },
+    {
+      "epoch": 0.091,
+      "eval_cos_sim": 0.8560004234313965,
+      "eval_loss": 0.14498550846325817,
+      "eval_runtime": 166.8423,
+      "eval_samples_per_second": 23.975,
+      "eval_steps_per_second": 0.378,
+      "step": 910
+    },
+    {
+      "epoch": 0.092,
+      "grad_norm": 0.17160160839557648,
+      "learning_rate": 4.600000000000001e-05,
+      "loss": 0.1387,
+      "step": 920
+    },
+    {
+      "epoch": 0.092,
+      "eval_cos_sim": 0.8561151623725891,
+      "eval_loss": 0.14487657501446666,
+      "eval_runtime": 166.4947,
+      "eval_samples_per_second": 24.025,
+      "eval_steps_per_second": 0.378,
+      "step": 920
+    },
+    {
+      "epoch": 0.093,
+      "grad_norm": 0.18101643025875092,
+      "learning_rate": 4.6500000000000005e-05,
+      "loss": 0.1362,
+      "step": 930
+    },
+    {
+      "epoch": 0.093,
+      "eval_cos_sim": 0.8565106987953186,
+      "eval_loss": 0.1444753079627699,
+      "eval_runtime": 162.5163,
+      "eval_samples_per_second": 24.613,
+      "eval_steps_per_second": 0.388,
+      "step": 930
+    },
+    {
+      "epoch": 0.094,
+      "grad_norm": 0.1913219839334488,
+      "learning_rate": 4.7e-05,
+      "loss": 0.137,
+      "step": 940
+    },
+    {
+      "epoch": 0.094,
+      "eval_cos_sim": 0.8572973608970642,
+      "eval_loss": 0.14369207908856332,
+      "eval_runtime": 159.6717,
+      "eval_samples_per_second": 25.051,
+      "eval_steps_per_second": 0.395,
+      "step": 940
+    },
+    {
+      "epoch": 0.095,
+      "grad_norm": 0.18858781456947327,
+      "learning_rate": 4.75e-05,
+      "loss": 0.1365,
+      "step": 950
+    },
+    {
+      "epoch": 0.095,
+      "eval_cos_sim": 0.8581430912017822,
+      "eval_loss": 0.1428443865989393,
+      "eval_runtime": 160.8669,
+      "eval_samples_per_second": 24.865,
+      "eval_steps_per_second": 0.392,
+      "step": 950
+    },
+    {
+      "epoch": 0.096,
+      "grad_norm": 0.1602417230606079,
+      "learning_rate": 4.8e-05,
+      "loss": 0.1353,
+      "step": 960
+    },
+    {
+      "epoch": 0.096,
+      "eval_cos_sim": 0.8581015467643738,
+      "eval_loss": 0.14288127615200938,
+      "eval_runtime": 164.3422,
+      "eval_samples_per_second": 24.339,
+      "eval_steps_per_second": 0.383,
+      "step": 960
+    },
+    {
+      "epoch": 0.097,
+      "grad_norm": 0.1659657508134842,
+      "learning_rate": 4.85e-05,
+      "loss": 0.1347,
+      "step": 970
+    },
+    {
+      "epoch": 0.097,
+      "eval_cos_sim": 0.8594202399253845,
+      "eval_loss": 0.14155777981030407,
+      "eval_runtime": 167.1461,
+      "eval_samples_per_second": 23.931,
+      "eval_steps_per_second": 0.377,
+      "step": 970
+    },
+    {
+      "epoch": 0.098,
+      "grad_norm": 0.21223703026771545,
+      "learning_rate": 4.9e-05,
+      "loss": 0.138,
+      "step": 980
+    },
+    {
+      "epoch": 0.098,
+      "eval_cos_sim": 0.8584538698196411,
+      "eval_loss": 0.14252794553982676,
+      "eval_runtime": 167.1973,
+      "eval_samples_per_second": 23.924,
+      "eval_steps_per_second": 0.377,
+      "step": 980
+    },
+    {
+      "epoch": 0.099,
+      "grad_norm": 0.15602050721645355,
+      "learning_rate": 4.9500000000000004e-05,
+      "loss": 0.1392,
+      "step": 990
+    },
+    {
+      "epoch": 0.099,
+      "eval_cos_sim": 0.859849750995636,
+      "eval_loss": 0.141124472162599,
+      "eval_runtime": 168.2536,
+      "eval_samples_per_second": 23.774,
+      "eval_steps_per_second": 0.374,
+      "step": 990
+    },
+    {
+      "epoch": 0.1,
+      "grad_norm": 0.1410510390996933,
+      "learning_rate": 5e-05,
+      "loss": 0.1359,
+      "step": 1000
+    },
+    {
+      "epoch": 0.1,
+      "eval_cos_sim": 0.8603095412254333,
+      "eval_loss": 0.14066851665722788,
+      "eval_runtime": 162.5936,
+      "eval_samples_per_second": 24.601,
+      "eval_steps_per_second": 0.387,
+      "step": 1000
+    },
+    {
+      "epoch": 0.101,
+      "grad_norm": 0.15686458349227905,
+      "learning_rate": 4.999984769144476e-05,
+      "loss": 0.1354,
+      "step": 1010
+    },
+    {
+      "epoch": 0.101,
+      "eval_cos_sim": 0.8605126142501831,
+      "eval_loss": 0.14045782043682994,
+      "eval_runtime": 164.632,
+      "eval_samples_per_second": 24.297,
+      "eval_steps_per_second": 0.383,
+      "step": 1010
+    },
+    {
+      "epoch": 0.102,
+      "grad_norm": 0.1402912735939026,
+      "learning_rate": 4.999939076763487e-05,
+      "loss": 0.1333,
+      "step": 1020
+    },
+    {
+      "epoch": 0.102,
+      "eval_cos_sim": 0.8615362644195557,
+      "eval_loss": 0.1394294276450819,
+      "eval_runtime": 165.0933,
+      "eval_samples_per_second": 24.229,
+      "eval_steps_per_second": 0.382,
+      "step": 1020
+    },
+    {
+      "epoch": 0.103,
+      "grad_norm": 0.14782707393169403,
+      "learning_rate": 4.999862923413781e-05,
+      "loss": 0.1354,
+      "step": 1030
+    },
+    {
+      "epoch": 0.103,
+      "eval_cos_sim": 0.8612933158874512,
+      "eval_loss": 0.13967935611950816,
+      "eval_runtime": 164.3708,
+      "eval_samples_per_second": 24.335,
+      "eval_steps_per_second": 0.383,
+      "step": 1030
+    },
+    {
+      "epoch": 0.104,
+      "grad_norm": 0.14607678353786469,
+      "learning_rate": 4.999756310023261e-05,
+      "loss": 0.1336,
+      "step": 1040
+    },
+    {
+      "epoch": 0.104,
+      "eval_cos_sim": 0.861626148223877,
+      "eval_loss": 0.13934949829327525,
+      "eval_runtime": 165.2839,
+      "eval_samples_per_second": 24.201,
+      "eval_steps_per_second": 0.381,
+      "step": 1040
+    },
+    {
+      "epoch": 0.105,
+      "grad_norm": 0.1559648960828781,
+      "learning_rate": 4.9996192378909786e-05,
+      "loss": 0.1382,
+      "step": 1050
+    },
+    {
+      "epoch": 0.105,
+      "eval_cos_sim": 0.8621127009391785,
+      "eval_loss": 0.13886528732525766,
+      "eval_runtime": 175.2773,
+      "eval_samples_per_second": 22.821,
+      "eval_steps_per_second": 0.359,
+      "step": 1050
+    },
+    {
+      "epoch": 0.106,
+      "grad_norm": 0.15351799130439758,
+      "learning_rate": 4.999451708687114e-05,
+      "loss": 0.133,
+      "step": 1060
+    },
+    {
+      "epoch": 0.106,
+      "eval_cos_sim": 0.8621622920036316,
+      "eval_loss": 0.1388200798247999,
+      "eval_runtime": 168.4264,
+      "eval_samples_per_second": 23.749,
+      "eval_steps_per_second": 0.374,
+      "step": 1060
+    },
+    {
+      "epoch": 0.107,
+      "grad_norm": 0.13820523023605347,
+      "learning_rate": 4.999253724452958e-05,
+      "loss": 0.1314,
+      "step": 1070
+    },
+    {
+      "epoch": 0.107,
+      "eval_cos_sim": 0.8628634810447693,
+      "eval_loss": 0.13812118484723032,
+      "eval_runtime": 160.5977,
+      "eval_samples_per_second": 24.907,
+      "eval_steps_per_second": 0.392,
+      "step": 1070
+    },
+    {
+      "epoch": 0.108,
+      "grad_norm": 0.1627858728170395,
+      "learning_rate": 4.999025287600886e-05,
+      "loss": 0.1327,
+      "step": 1080
+    },
+    {
+      "epoch": 0.108,
+      "eval_cos_sim": 0.8630489706993103,
+      "eval_loss": 0.13793376447903574,
+      "eval_runtime": 165.8793,
+      "eval_samples_per_second": 24.114,
+      "eval_steps_per_second": 0.38,
+      "step": 1080
+    },
+    {
+      "epoch": 0.109,
+      "grad_norm": 0.14537328481674194,
+      "learning_rate": 4.998766400914329e-05,
+      "loss": 0.1347,
+      "step": 1090
+    },
+    {
+      "epoch": 0.109,
+      "eval_cos_sim": 0.8635595440864563,
+      "eval_loss": 0.13742092325436533,
+      "eval_runtime": 175.092,
+      "eval_samples_per_second": 22.845,
+      "eval_steps_per_second": 0.36,
+      "step": 1090
+    },
+    {
+      "epoch": 0.11,
+      "grad_norm": 0.15490968525409698,
+      "learning_rate": 4.99847706754774e-05,
+      "loss": 0.1322,
+      "step": 1100
+    },
+    {
+      "epoch": 0.11,
+      "eval_cos_sim": 0.8640041351318359,
+      "eval_loss": 0.13697428562390268,
+      "eval_runtime": 171.1077,
+      "eval_samples_per_second": 23.377,
+      "eval_steps_per_second": 0.368,
+      "step": 1100
+    },
+    {
+      "epoch": 0.111,
+      "grad_norm": 0.13356918096542358,
+      "learning_rate": 4.998157291026553e-05,
+      "loss": 0.1346,
+      "step": 1110
+    },
+    {
+      "epoch": 0.111,
+      "eval_cos_sim": 0.864433228969574,
+      "eval_loss": 0.13654097273098886,
+      "eval_runtime": 165.0965,
+      "eval_samples_per_second": 24.228,
+      "eval_steps_per_second": 0.382,
+      "step": 1110
+    },
+    {
+      "epoch": 0.112,
+      "grad_norm": 0.1531037539243698,
+      "learning_rate": 4.997807075247146e-05,
+      "loss": 0.131,
+      "step": 1120
+    },
+    {
+      "epoch": 0.112,
+      "eval_cos_sim": 0.8644208908081055,
+      "eval_loss": 0.13655049898373545,
+      "eval_runtime": 175.2585,
+      "eval_samples_per_second": 22.823,
+      "eval_steps_per_second": 0.359,
+      "step": 1120
+    },
+    {
+      "epoch": 0.113,
+      "grad_norm": 0.14743933081626892,
+      "learning_rate": 4.997426424476787e-05,
+      "loss": 0.131,
+      "step": 1130
+    },
+    {
+      "epoch": 0.113,
+      "eval_cos_sim": 0.8642333745956421,
+      "eval_loss": 0.1367411518310255,
+      "eval_runtime": 175.2554,
+      "eval_samples_per_second": 22.824,
+      "eval_steps_per_second": 0.359,
+      "step": 1130
+    },
+    {
+      "epoch": 0.114,
+      "grad_norm": 0.15171775221824646,
+      "learning_rate": 4.997015343353585e-05,
+      "loss": 0.1288,
+      "step": 1140
+    },
+    {
+      "epoch": 0.114,
+      "eval_cos_sim": 0.864815890789032,
+      "eval_loss": 0.13617111923443737,
+      "eval_runtime": 168.6886,
+      "eval_samples_per_second": 23.712,
+      "eval_steps_per_second": 0.373,
+      "step": 1140
+    },
+    {
+      "epoch": 0.115,
+      "grad_norm": 0.17678213119506836,
+      "learning_rate": 4.996573836886435e-05,
+      "loss": 0.1318,
+      "step": 1150
+    },
+    {
+      "epoch": 0.115,
+      "eval_cos_sim": 0.865388035774231,
+      "eval_loss": 0.13559978248822155,
+      "eval_runtime": 164.5713,
+      "eval_samples_per_second": 24.306,
+      "eval_steps_per_second": 0.383,
+      "step": 1150
+    },
+    {
+      "epoch": 0.116,
+      "grad_norm": 0.19686037302017212,
+      "learning_rate": 4.996101910454953e-05,
+      "loss": 0.1328,
+      "step": 1160
+    },
+    {
+      "epoch": 0.116,
+      "eval_cos_sim": 0.8653638362884521,
+      "eval_loss": 0.13562491752850475,
+      "eval_runtime": 162.5909,
+      "eval_samples_per_second": 24.602,
+      "eval_steps_per_second": 0.387,
+      "step": 1160
+    },
+    {
+      "epoch": 0.117,
+      "grad_norm": 0.13033832609653473,
+      "learning_rate": 4.995599569809414e-05,
+      "loss": 0.1314,
+      "step": 1170
+    },
+    {
+      "epoch": 0.117,
+      "eval_cos_sim": 0.8652835488319397,
+      "eval_loss": 0.1357041087363905,
+      "eval_runtime": 166.4639,
+      "eval_samples_per_second": 24.029,
+      "eval_steps_per_second": 0.378,
+      "step": 1170
+    },
+    {
+      "epoch": 0.118,
+      "grad_norm": 0.14916931092739105,
+      "learning_rate": 4.995066821070679e-05,
+      "loss": 0.1313,
+      "step": 1180
+    },
+    {
+      "epoch": 0.118,
+      "eval_cos_sim": 0.8656159043312073,
+      "eval_loss": 0.13536836673962535,
+      "eval_runtime": 162.1775,
+      "eval_samples_per_second": 24.664,
+      "eval_steps_per_second": 0.388,
+      "step": 1180
+    },
+    {
+      "epoch": 0.119,
+      "grad_norm": 0.14063435792922974,
+      "learning_rate": 4.994503670730125e-05,
+      "loss": 0.1292,
+      "step": 1190
+    },
+    {
+      "epoch": 0.119,
+      "eval_cos_sim": 0.8658972978591919,
+      "eval_loss": 0.13509506275402963,
+      "eval_runtime": 163.733,
+      "eval_samples_per_second": 24.43,
+      "eval_steps_per_second": 0.385,
+      "step": 1190
+    },
+    {
+      "epoch": 0.12,
+      "grad_norm": 0.12632516026496887,
+      "learning_rate": 4.993910125649561e-05,
+      "loss": 0.1291,
+      "step": 1200
+    },
+    {
+      "epoch": 0.12,
+      "eval_cos_sim": 0.8662503361701965,
+      "eval_loss": 0.13473713161694467,
+      "eval_runtime": 165.846,
+      "eval_samples_per_second": 24.119,
+      "eval_steps_per_second": 0.38,
+      "step": 1200
+    },
+    {
+      "epoch": 0.121,
+      "grad_norm": 0.1310836374759674,
+      "learning_rate": 4.9932861930611454e-05,
+      "loss": 0.1323,
+      "step": 1210
+    },
+    {
+      "epoch": 0.121,
+      "eval_cos_sim": 0.8662631511688232,
+      "eval_loss": 0.13472085668789804,
+      "eval_runtime": 165.8256,
+      "eval_samples_per_second": 24.122,
+      "eval_steps_per_second": 0.38,
+      "step": 1210
+    },
+    {
+      "epoch": 0.122,
+      "grad_norm": 0.14743076264858246,
+      "learning_rate": 4.992631880567301e-05,
+      "loss": 0.1296,
+      "step": 1220
+    },
+    {
+      "epoch": 0.122,
+      "eval_cos_sim": 0.8664645552635193,
+      "eval_loss": 0.13453043796765268,
+      "eval_runtime": 161.1688,
+      "eval_samples_per_second": 24.819,
+      "eval_steps_per_second": 0.391,
+      "step": 1220
+    },
+    {
+      "epoch": 0.123,
+      "grad_norm": 0.14941203594207764,
+      "learning_rate": 4.991947196140618e-05,
+      "loss": 0.127,
+      "step": 1230
+    },
+    {
+      "epoch": 0.123,
+      "eval_cos_sim": 0.8669182658195496,
+      "eval_loss": 0.1340716586326307,
+      "eval_runtime": 159.0609,
+      "eval_samples_per_second": 25.148,
+      "eval_steps_per_second": 0.396,
+      "step": 1230
+    },
+    {
+      "epoch": 0.124,
+      "grad_norm": 0.15963555872440338,
+      "learning_rate": 4.991232148123761e-05,
+      "loss": 0.1286,
+      "step": 1240
+    },
+    {
+      "epoch": 0.124,
+      "eval_cos_sim": 0.8672899007797241,
+      "eval_loss": 0.13370788147198617,
+      "eval_runtime": 162.1984,
+      "eval_samples_per_second": 24.661,
+      "eval_steps_per_second": 0.388,
+      "step": 1240
+    },
+    {
+      "epoch": 0.125,
+      "grad_norm": 0.12001099437475204,
+      "learning_rate": 4.990486745229364e-05,
+      "loss": 0.1289,
+      "step": 1250
+    },
+    {
+      "epoch": 0.125,
+      "eval_cos_sim": 0.8670727610588074,
+      "eval_loss": 0.13392866852032603,
+      "eval_runtime": 161.0006,
+      "eval_samples_per_second": 24.845,
+      "eval_steps_per_second": 0.391,
+      "step": 1250
+    },
+    {
+      "epoch": 0.126,
+      "grad_norm": 0.13203372061252594,
+      "learning_rate": 4.989710996539926e-05,
+      "loss": 0.1268,
+      "step": 1260
+    },
+    {
+      "epoch": 0.126,
+      "eval_cos_sim": 0.8671303987503052,
+      "eval_loss": 0.1338693347190565,
+      "eval_runtime": 164.8756,
+      "eval_samples_per_second": 24.261,
+      "eval_steps_per_second": 0.382,
+      "step": 1260
+    },
+    {
+      "epoch": 0.127,
+      "grad_norm": 0.1145872101187706,
+      "learning_rate": 4.9889049115077005e-05,
+      "loss": 0.1274,
+      "step": 1270
+    },
+    {
+      "epoch": 0.127,
+      "eval_cos_sim": 0.8675019145011902,
+      "eval_loss": 0.13349107553708017,
+      "eval_runtime": 169.3092,
+      "eval_samples_per_second": 23.625,
+      "eval_steps_per_second": 0.372,
+      "step": 1270
+    },
+    {
+      "epoch": 0.128,
+      "grad_norm": 0.1377398520708084,
+      "learning_rate": 4.988068499954578e-05,
+      "loss": 0.1317,
+      "step": 1280
+    },
+    {
+      "epoch": 0.128,
+      "eval_cos_sim": 0.8676916360855103,
+      "eval_loss": 0.1333048849319166,
+      "eval_runtime": 159.5371,
+      "eval_samples_per_second": 25.073,
+      "eval_steps_per_second": 0.395,
+      "step": 1280
+    },
+    {
+      "epoch": 0.129,
+      "grad_norm": 0.170301154255867,
+      "learning_rate": 4.987201772071971e-05,
+      "loss": 0.1285,
+      "step": 1290
+    },
+    {
+      "epoch": 0.129,
+      "eval_cos_sim": 0.8680738806724548,
+      "eval_loss": 0.13292226507412852,
+      "eval_runtime": 176.1386,
+      "eval_samples_per_second": 22.709,
+      "eval_steps_per_second": 0.358,
+      "step": 1290
+    },
+    {
+      "epoch": 0.13,
+      "grad_norm": 0.1578381359577179,
+      "learning_rate": 4.9863047384206835e-05,
+      "loss": 0.1295,
+      "step": 1300
+    },
+    {
+      "epoch": 0.13,
+      "eval_cos_sim": 0.8683300614356995,
+      "eval_loss": 0.13265489532696664,
+      "eval_runtime": 164.7229,
+      "eval_samples_per_second": 24.283,
+      "eval_steps_per_second": 0.382,
+      "step": 1300
+    },
+    {
+      "epoch": 0.131,
+      "grad_norm": 0.10768606513738632,
+      "learning_rate": 4.985377409930789e-05,
+      "loss": 0.132,
+      "step": 1310
+    },
+    {
+      "epoch": 0.131,
+      "eval_cos_sim": 0.8683825731277466,
+      "eval_loss": 0.13260123255001963,
+      "eval_runtime": 163.9895,
+      "eval_samples_per_second": 24.392,
+      "eval_steps_per_second": 0.384,
+      "step": 1310
+    },
+    {
+      "epoch": 0.132,
+      "grad_norm": 0.15251098573207855,
+      "learning_rate": 4.984419797901491e-05,
+      "loss": 0.128,
+      "step": 1320
+    },
+    {
+      "epoch": 0.132,
+      "eval_cos_sim": 0.868379533290863,
+      "eval_loss": 0.1326023030494398,
+      "eval_runtime": 160.1598,
+      "eval_samples_per_second": 24.975,
+      "eval_steps_per_second": 0.393,
+      "step": 1320
+    },
+    {
+      "epoch": 0.133,
+      "grad_norm": 0.1114293709397316,
+      "learning_rate": 4.983431914000991e-05,
+      "loss": 0.1293,
+      "step": 1330
+    },
+    {
+      "epoch": 0.133,
+      "eval_cos_sim": 0.8686286211013794,
+      "eval_loss": 0.13235191013562145,
+      "eval_runtime": 168.1325,
+      "eval_samples_per_second": 23.791,
+      "eval_steps_per_second": 0.375,
+      "step": 1330
+    },
+    {
+      "epoch": 0.134,
+      "grad_norm": 0.11598256975412369,
+      "learning_rate": 4.982413770266342e-05,
+      "loss": 0.1321,
+      "step": 1340
+    },
+    {
+      "epoch": 0.134,
+      "eval_cos_sim": 0.8685267567634583,
+      "eval_loss": 0.1324642758582777,
+      "eval_runtime": 168.2142,
+      "eval_samples_per_second": 23.779,
+      "eval_steps_per_second": 0.375,
+      "step": 1340
+    },
+    {
+      "epoch": 0.135,
+      "grad_norm": 0.17224518954753876,
+      "learning_rate": 4.9813653791033057e-05,
+      "loss": 0.1293,
+      "step": 1350
+    },
+    {
+      "epoch": 0.135,
+      "eval_cos_sim": 0.8683568239212036,
+      "eval_loss": 0.13261960174786508,
+      "eval_runtime": 160.932,
+      "eval_samples_per_second": 24.855,
+      "eval_steps_per_second": 0.391,
+      "step": 1350
+    },
+    {
+      "epoch": 0.136,
+      "grad_norm": 0.10910200327634811,
+      "learning_rate": 4.980286753286195e-05,
+      "loss": 0.1279,
+      "step": 1360
+    },
+    {
+      "epoch": 0.136,
+      "eval_cos_sim": 0.8685951828956604,
+      "eval_loss": 0.13237247326123178,
+      "eval_runtime": 160.5826,
+      "eval_samples_per_second": 24.909,
+      "eval_steps_per_second": 0.392,
+      "step": 1360
+    },
+    {
+      "epoch": 0.137,
+      "grad_norm": 0.12568902969360352,
+      "learning_rate": 4.979177905957726e-05,
+      "loss": 0.1276,
+      "step": 1370
+    },
+    {
+      "epoch": 0.137,
+      "eval_cos_sim": 0.8692377209663391,
+      "eval_loss": 0.1317250809882826,
+      "eval_runtime": 162.4444,
+      "eval_samples_per_second": 24.624,
+      "eval_steps_per_second": 0.388,
+      "step": 1370
+    },
+    {
+      "epoch": 0.138,
+      "grad_norm": 0.13634935021400452,
+      "learning_rate": 4.978038850628854e-05,
+      "loss": 0.1287,
+      "step": 1380
+    },
+    {
+      "epoch": 0.138,
+      "eval_cos_sim": 0.8693557381629944,
+      "eval_loss": 0.13161035587536757,
+      "eval_runtime": 167.3015,
+      "eval_samples_per_second": 23.909,
+      "eval_steps_per_second": 0.377,
+      "step": 1380
+    },
+    {
+      "epoch": 0.139,
+      "grad_norm": 0.10860959440469742,
+      "learning_rate": 4.976869601178609e-05,
+      "loss": 0.1263,
+      "step": 1390
+    },
+    {
+      "epoch": 0.139,
+      "eval_cos_sim": 0.8695587515830994,
+      "eval_loss": 0.13140406753766004,
+      "eval_runtime": 185.1037,
+      "eval_samples_per_second": 21.61,
+      "eval_steps_per_second": 0.34,
+      "step": 1390
+    },
+    {
+      "epoch": 0.14,
+      "grad_norm": 0.10750249028205872,
+      "learning_rate": 4.975670171853926e-05,
+      "loss": 0.1262,
+      "step": 1400
+    },
+    {
+      "epoch": 0.14,
+      "eval_cos_sim": 0.8693708181381226,
+      "eval_loss": 0.13159809972989026,
+      "eval_runtime": 194.8811,
+      "eval_samples_per_second": 20.525,
+      "eval_steps_per_second": 0.323,
+      "step": 1400
+    }
+  ],
+  "logging_steps": 10,
+  "max_steps": 10000,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 9223372036854775807,
+  "save_steps": 10,
+  "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": 160,
+  "trial_name": null,
+  "trial_params": null
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:228943860476aef422f429bf676bbb7f95e79b4b2c6e641a22176011b8628860
+size 5176