Training in progress, step 282

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:f13d2d2ee4a89f12eb24c2531e16499da0ada8ea331ec110dbd806519b8eb9b4
+oid sha256:0263202fc195d3f84616e80a536b52506dab04b8e4e367bb1b9bcfbd400f3101
 size 605156676

run-bq1dw9ny/checkpoint-282/config.json

@@ -0,0 +1,22 @@
+{
+  "_name_or_path": "openai/clip-vit-base-patch32",
+  "architectures": [
+    "CLIPModel"
+  ],
+  "initializer_factor": 1.0,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "clip",
+  "projection_dim": 512,
+  "text_config": {
+    "bos_token_id": 0,
+    "dropout": 0.0,
+    "eos_token_id": 2,
+    "model_type": "clip_text_model"
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.42.0.dev0",
+  "vision_config": {
+    "dropout": 0.0,
+    "model_type": "clip_vision_model"
+  }
+}

run-bq1dw9ny/checkpoint-282/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0263202fc195d3f84616e80a536b52506dab04b8e4e367bb1b9bcfbd400f3101
+size 605156676

run-bq1dw9ny/checkpoint-282/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7b01104df926457f0eeb1642e7e785485bc4fc5993053aead19db11f58ee503a
+size 1210551612

run-bq1dw9ny/checkpoint-282/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d5c33740806ccf1955d4345523f1b05c54a20fdcaf53bdbe6b546e984ef823f1
+size 14244

run-bq1dw9ny/checkpoint-282/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f867e47eeac9ea634f86ff11b0978933721dd3b3057b39790e98c44aab4c5a70
+size 1064

run-bq1dw9ny/checkpoint-282/trainer_state.json

@@ -0,0 +1,2155 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 3.0,
+  "eval_steps": 2,
+  "global_step": 282,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.02127659574468085,
+      "grad_norm": 147.76431274414062,
+      "learning_rate": 0.000495960190363068,
+      "loss": 3.2762,
+      "step": 2
+    },
+    {
+      "epoch": 0.02127659574468085,
+      "eval_loss": 5.192628383636475,
+      "eval_runtime": 254.7582,
+      "eval_samples_per_second": 1.178,
+      "eval_steps_per_second": 0.02,
+      "step": 2
+    },
+    {
+      "epoch": 0.0425531914893617,
+      "grad_norm": 34.33057403564453,
+      "learning_rate": 0.0004924176175747603,
+      "loss": 4.1983,
+      "step": 4
+    },
+    {
+      "epoch": 0.0425531914893617,
+      "eval_loss": 4.506951332092285,
+      "eval_runtime": 116.2019,
+      "eval_samples_per_second": 2.582,
+      "eval_steps_per_second": 0.043,
+      "step": 4
+    },
+    {
+      "epoch": 0.06382978723404255,
+      "grad_norm": 2.5619542598724365,
+      "learning_rate": 0.0004888750447864527,
+      "loss": 3.6907,
+      "step": 6
+    },
+    {
+      "epoch": 0.06382978723404255,
+      "eval_loss": 4.125861644744873,
+      "eval_runtime": 95.9708,
+      "eval_samples_per_second": 3.126,
+      "eval_steps_per_second": 0.052,
+      "step": 6
+    },
+    {
+      "epoch": 0.0851063829787234,
+      "grad_norm": 1.6348135471343994,
+      "learning_rate": 0.0004853324719981451,
+      "loss": 3.5109,
+      "step": 8
+    },
+    {
+      "epoch": 0.0851063829787234,
+      "eval_loss": 4.108506679534912,
+      "eval_runtime": 95.4994,
+      "eval_samples_per_second": 3.141,
+      "eval_steps_per_second": 0.052,
+      "step": 8
+    },
+    {
+      "epoch": 0.10638297872340426,
+      "grad_norm": 0.02139226719737053,
+      "learning_rate": 0.00048178989920983744,
+      "loss": 3.4686,
+      "step": 10
+    },
+    {
+      "epoch": 0.10638297872340426,
+      "eval_loss": 4.104280471801758,
+      "eval_runtime": 114.204,
+      "eval_samples_per_second": 2.627,
+      "eval_steps_per_second": 0.044,
+      "step": 10
+    },
+    {
+      "epoch": 0.1276595744680851,
+      "grad_norm": 0.005519088823348284,
+      "learning_rate": 0.0004782473264215299,
+      "loss": 3.466,
+      "step": 12
+    },
+    {
+      "epoch": 0.1276595744680851,
+      "eval_loss": 4.104023456573486,
+      "eval_runtime": 96.7284,
+      "eval_samples_per_second": 3.101,
+      "eval_steps_per_second": 0.052,
+      "step": 12
+    },
+    {
+      "epoch": 0.14893617021276595,
+      "grad_norm": 0.005563751328736544,
+      "learning_rate": 0.00047470475363322223,
+      "loss": 3.4658,
+      "step": 14
+    },
+    {
+      "epoch": 0.14893617021276595,
+      "eval_loss": 4.104072093963623,
+      "eval_runtime": 97.0883,
+      "eval_samples_per_second": 3.09,
+      "eval_steps_per_second": 0.051,
+      "step": 14
+    },
+    {
+      "epoch": 0.1702127659574468,
+      "grad_norm": 0.0055496166460216045,
+      "learning_rate": 0.00047116218084491463,
+      "loss": 3.4659,
+      "step": 16
+    },
+    {
+      "epoch": 0.1702127659574468,
+      "eval_loss": 4.104039669036865,
+      "eval_runtime": 93.8461,
+      "eval_samples_per_second": 3.197,
+      "eval_steps_per_second": 0.053,
+      "step": 16
+    },
+    {
+      "epoch": 0.19148936170212766,
+      "grad_norm": 0.035093989223241806,
+      "learning_rate": 0.000467619608056607,
+      "loss": 3.4659,
+      "step": 18
+    },
+    {
+      "epoch": 0.19148936170212766,
+      "eval_loss": 4.104025363922119,
+      "eval_runtime": 93.5115,
+      "eval_samples_per_second": 3.208,
+      "eval_steps_per_second": 0.053,
+      "step": 18
+    },
+    {
+      "epoch": 0.2127659574468085,
+      "grad_norm": 0.013235281221568584,
+      "learning_rate": 0.0004640770352682993,
+      "loss": 3.4658,
+      "step": 20
+    },
+    {
+      "epoch": 0.2127659574468085,
+      "eval_loss": 4.1039652824401855,
+      "eval_runtime": 93.8067,
+      "eval_samples_per_second": 3.198,
+      "eval_steps_per_second": 0.053,
+      "step": 20
+    },
+    {
+      "epoch": 0.23404255319148937,
+      "grad_norm": 0.0016526976833119988,
+      "learning_rate": 0.0004605344624799917,
+      "loss": 3.4658,
+      "step": 22
+    },
+    {
+      "epoch": 0.23404255319148937,
+      "eval_loss": 4.103963375091553,
+      "eval_runtime": 93.4791,
+      "eval_samples_per_second": 3.209,
+      "eval_steps_per_second": 0.053,
+      "step": 22
+    },
+    {
+      "epoch": 0.2553191489361702,
+      "grad_norm": 0.0017120653064921498,
+      "learning_rate": 0.00045699188969168406,
+      "loss": 3.4658,
+      "step": 24
+    },
+    {
+      "epoch": 0.2553191489361702,
+      "eval_loss": 4.103957176208496,
+      "eval_runtime": 116.2864,
+      "eval_samples_per_second": 2.58,
+      "eval_steps_per_second": 0.043,
+      "step": 24
+    },
+    {
+      "epoch": 0.2765957446808511,
+      "grad_norm": 0.0011341843055561185,
+      "learning_rate": 0.00045344931690337645,
+      "loss": 3.4658,
+      "step": 26
+    },
+    {
+      "epoch": 0.2765957446808511,
+      "eval_loss": 4.103954315185547,
+      "eval_runtime": 96.3574,
+      "eval_samples_per_second": 3.113,
+      "eval_steps_per_second": 0.052,
+      "step": 26
+    },
+    {
+      "epoch": 0.2978723404255319,
+      "grad_norm": 0.0025584695395082235,
+      "learning_rate": 0.0004499067441150688,
+      "loss": 3.4658,
+      "step": 28
+    },
+    {
+      "epoch": 0.2978723404255319,
+      "eval_loss": 4.103954792022705,
+      "eval_runtime": 109.8402,
+      "eval_samples_per_second": 2.731,
+      "eval_steps_per_second": 0.046,
+      "step": 28
+    },
+    {
+      "epoch": 0.3191489361702128,
+      "grad_norm": 0.0012942380271852016,
+      "learning_rate": 0.0004463641713267612,
+      "loss": 3.4658,
+      "step": 30
+    },
+    {
+      "epoch": 0.3191489361702128,
+      "eval_loss": 4.1039557456970215,
+      "eval_runtime": 114.0952,
+      "eval_samples_per_second": 2.629,
+      "eval_steps_per_second": 0.044,
+      "step": 30
+    },
+    {
+      "epoch": 0.3404255319148936,
+      "grad_norm": 0.001233804621733725,
+      "learning_rate": 0.00044282159853845354,
+      "loss": 3.4658,
+      "step": 32
+    },
+    {
+      "epoch": 0.3404255319148936,
+      "eval_loss": 4.103955268859863,
+      "eval_runtime": 96.1239,
+      "eval_samples_per_second": 3.121,
+      "eval_steps_per_second": 0.052,
+      "step": 32
+    },
+    {
+      "epoch": 0.3617021276595745,
+      "grad_norm": 0.0013235628139227629,
+      "learning_rate": 0.00043927902575014593,
+      "loss": 3.4658,
+      "step": 34
+    },
+    {
+      "epoch": 0.3617021276595745,
+      "eval_loss": 4.103952407836914,
+      "eval_runtime": 135.012,
+      "eval_samples_per_second": 2.222,
+      "eval_steps_per_second": 0.037,
+      "step": 34
+    },
+    {
+      "epoch": 0.3829787234042553,
+      "grad_norm": 0.002040724502876401,
+      "learning_rate": 0.00043573645296183833,
+      "loss": 3.4658,
+      "step": 36
+    },
+    {
+      "epoch": 0.3829787234042553,
+      "eval_loss": 4.103944301605225,
+      "eval_runtime": 96.6912,
+      "eval_samples_per_second": 3.103,
+      "eval_steps_per_second": 0.052,
+      "step": 36
+    },
+    {
+      "epoch": 0.40425531914893614,
+      "grad_norm": 0.0005005718558095396,
+      "learning_rate": 0.00043219388017353067,
+      "loss": 3.4658,
+      "step": 38
+    },
+    {
+      "epoch": 0.40425531914893614,
+      "eval_loss": 4.103936195373535,
+      "eval_runtime": 120.942,
+      "eval_samples_per_second": 2.481,
+      "eval_steps_per_second": 0.041,
+      "step": 38
+    },
+    {
+      "epoch": 0.425531914893617,
+      "grad_norm": 0.0004194887587800622,
+      "learning_rate": 0.00042865130738522307,
+      "loss": 3.4657,
+      "step": 40
+    },
+    {
+      "epoch": 0.425531914893617,
+      "eval_loss": 4.103933334350586,
+      "eval_runtime": 165.7391,
+      "eval_samples_per_second": 1.81,
+      "eval_steps_per_second": 0.03,
+      "step": 40
+    },
+    {
+      "epoch": 0.44680851063829785,
+      "grad_norm": 0.0007025453960523009,
+      "learning_rate": 0.0004251087345969154,
+      "loss": 3.4657,
+      "step": 42
+    },
+    {
+      "epoch": 0.44680851063829785,
+      "eval_loss": 4.103935241699219,
+      "eval_runtime": 95.9616,
+      "eval_samples_per_second": 3.126,
+      "eval_steps_per_second": 0.052,
+      "step": 42
+    },
+    {
+      "epoch": 0.46808510638297873,
+      "grad_norm": 0.0005715902079828084,
+      "learning_rate": 0.0004215661618086078,
+      "loss": 3.4657,
+      "step": 44
+    },
+    {
+      "epoch": 0.46808510638297873,
+      "eval_loss": 4.103936195373535,
+      "eval_runtime": 95.6382,
+      "eval_samples_per_second": 3.137,
+      "eval_steps_per_second": 0.052,
+      "step": 44
+    },
+    {
+      "epoch": 0.48936170212765956,
+      "grad_norm": 0.0009604791412129998,
+      "learning_rate": 0.00041802358902030015,
+      "loss": 3.4657,
+      "step": 46
+    },
+    {
+      "epoch": 0.48936170212765956,
+      "eval_loss": 4.103936195373535,
+      "eval_runtime": 95.422,
+      "eval_samples_per_second": 3.144,
+      "eval_steps_per_second": 0.052,
+      "step": 46
+    },
+    {
+      "epoch": 0.5106382978723404,
+      "grad_norm": 0.0015922917518764734,
+      "learning_rate": 0.00041448101623199255,
+      "loss": 3.4657,
+      "step": 48
+    },
+    {
+      "epoch": 0.5106382978723404,
+      "eval_loss": 4.103933334350586,
+      "eval_runtime": 95.2253,
+      "eval_samples_per_second": 3.15,
+      "eval_steps_per_second": 0.053,
+      "step": 48
+    },
+    {
+      "epoch": 0.5319148936170213,
+      "grad_norm": 0.00046437734272331,
+      "learning_rate": 0.0004109384434436849,
+      "loss": 3.4657,
+      "step": 50
+    },
+    {
+      "epoch": 0.5319148936170213,
+      "eval_loss": 4.103931427001953,
+      "eval_runtime": 93.7871,
+      "eval_samples_per_second": 3.199,
+      "eval_steps_per_second": 0.053,
+      "step": 50
+    },
+    {
+      "epoch": 0.5531914893617021,
+      "grad_norm": 0.0003185276291333139,
+      "learning_rate": 0.00040739587065537723,
+      "loss": 3.4657,
+      "step": 52
+    },
+    {
+      "epoch": 0.5531914893617021,
+      "eval_loss": 4.103931903839111,
+      "eval_runtime": 94.4747,
+      "eval_samples_per_second": 3.175,
+      "eval_steps_per_second": 0.053,
+      "step": 52
+    },
+    {
+      "epoch": 0.574468085106383,
+      "grad_norm": 0.0005938044050708413,
+      "learning_rate": 0.00040385329786706963,
+      "loss": 3.4657,
+      "step": 54
+    },
+    {
+      "epoch": 0.574468085106383,
+      "eval_loss": 4.103933334350586,
+      "eval_runtime": 94.4693,
+      "eval_samples_per_second": 3.176,
+      "eval_steps_per_second": 0.053,
+      "step": 54
+    },
+    {
+      "epoch": 0.5957446808510638,
+      "grad_norm": 0.0008093913784250617,
+      "learning_rate": 0.000400310725078762,
+      "loss": 3.4657,
+      "step": 56
+    },
+    {
+      "epoch": 0.5957446808510638,
+      "eval_loss": 4.103933334350586,
+      "eval_runtime": 93.6969,
+      "eval_samples_per_second": 3.202,
+      "eval_steps_per_second": 0.053,
+      "step": 56
+    },
+    {
+      "epoch": 0.6170212765957447,
+      "grad_norm": 0.000580158899538219,
+      "learning_rate": 0.0003967681522904544,
+      "loss": 3.4657,
+      "step": 58
+    },
+    {
+      "epoch": 0.6170212765957447,
+      "eval_loss": 4.103932857513428,
+      "eval_runtime": 94.0277,
+      "eval_samples_per_second": 3.191,
+      "eval_steps_per_second": 0.053,
+      "step": 58
+    },
+    {
+      "epoch": 0.6382978723404256,
+      "grad_norm": 0.0005565917235799134,
+      "learning_rate": 0.00039322557950214677,
+      "loss": 3.4657,
+      "step": 60
+    },
+    {
+      "epoch": 0.6382978723404256,
+      "eval_loss": 4.103931903839111,
+      "eval_runtime": 107.4438,
+      "eval_samples_per_second": 2.792,
+      "eval_steps_per_second": 0.047,
+      "step": 60
+    },
+    {
+      "epoch": 0.6595744680851063,
+      "grad_norm": 0.0006027113413438201,
+      "learning_rate": 0.00038968300671383916,
+      "loss": 3.4657,
+      "step": 62
+    },
+    {
+      "epoch": 0.6595744680851063,
+      "eval_loss": 4.103930473327637,
+      "eval_runtime": 107.2976,
+      "eval_samples_per_second": 2.796,
+      "eval_steps_per_second": 0.047,
+      "step": 62
+    },
+    {
+      "epoch": 0.6808510638297872,
+      "grad_norm": 0.0002680857141967863,
+      "learning_rate": 0.0003861404339255315,
+      "loss": 3.4657,
+      "step": 64
+    },
+    {
+      "epoch": 0.6808510638297872,
+      "eval_loss": 4.103930473327637,
+      "eval_runtime": 94.4753,
+      "eval_samples_per_second": 3.175,
+      "eval_steps_per_second": 0.053,
+      "step": 64
+    },
+    {
+      "epoch": 0.7021276595744681,
+      "grad_norm": 0.0003917052235919982,
+      "learning_rate": 0.0003825978611372239,
+      "loss": 3.4657,
+      "step": 66
+    },
+    {
+      "epoch": 0.7021276595744681,
+      "eval_loss": 4.103930473327637,
+      "eval_runtime": 94.4775,
+      "eval_samples_per_second": 3.175,
+      "eval_steps_per_second": 0.053,
+      "step": 66
+    },
+    {
+      "epoch": 0.723404255319149,
+      "grad_norm": 0.00032164924778044224,
+      "learning_rate": 0.00037905528834891625,
+      "loss": 3.4657,
+      "step": 68
+    },
+    {
+      "epoch": 0.723404255319149,
+      "eval_loss": 4.103930473327637,
+      "eval_runtime": 94.9268,
+      "eval_samples_per_second": 3.16,
+      "eval_steps_per_second": 0.053,
+      "step": 68
+    },
+    {
+      "epoch": 0.7446808510638298,
+      "grad_norm": 0.0004127752035856247,
+      "learning_rate": 0.0003755127155606086,
+      "loss": 3.4657,
+      "step": 70
+    },
+    {
+      "epoch": 0.7446808510638298,
+      "eval_loss": 4.103930473327637,
+      "eval_runtime": 95.0978,
+      "eval_samples_per_second": 3.155,
+      "eval_steps_per_second": 0.053,
+      "step": 70
+    },
+    {
+      "epoch": 0.7659574468085106,
+      "grad_norm": 0.00039929302874952555,
+      "learning_rate": 0.000371970142772301,
+      "loss": 3.4657,
+      "step": 72
+    },
+    {
+      "epoch": 0.7659574468085106,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 100.3564,
+      "eval_samples_per_second": 2.989,
+      "eval_steps_per_second": 0.05,
+      "step": 72
+    },
+    {
+      "epoch": 0.7872340425531915,
+      "grad_norm": 0.00019805252668447793,
+      "learning_rate": 0.00036842756998399333,
+      "loss": 3.4657,
+      "step": 74
+    },
+    {
+      "epoch": 0.7872340425531915,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 93.9211,
+      "eval_samples_per_second": 3.194,
+      "eval_steps_per_second": 0.053,
+      "step": 74
+    },
+    {
+      "epoch": 0.8085106382978723,
+      "grad_norm": 0.00028355230460874736,
+      "learning_rate": 0.0003648849971956858,
+      "loss": 3.4657,
+      "step": 76
+    },
+    {
+      "epoch": 0.8085106382978723,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 122.794,
+      "eval_samples_per_second": 2.443,
+      "eval_steps_per_second": 0.041,
+      "step": 76
+    },
+    {
+      "epoch": 0.8297872340425532,
+      "grad_norm": 0.0003330525360070169,
+      "learning_rate": 0.0003613424244073781,
+      "loss": 3.4657,
+      "step": 78
+    },
+    {
+      "epoch": 0.8297872340425532,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 108.2578,
+      "eval_samples_per_second": 2.771,
+      "eval_steps_per_second": 0.046,
+      "step": 78
+    },
+    {
+      "epoch": 0.851063829787234,
+      "grad_norm": 0.00028144015232101083,
+      "learning_rate": 0.0003577998516190705,
+      "loss": 3.4657,
+      "step": 80
+    },
+    {
+      "epoch": 0.851063829787234,
+      "eval_loss": 4.10392951965332,
+      "eval_runtime": 93.7754,
+      "eval_samples_per_second": 3.199,
+      "eval_steps_per_second": 0.053,
+      "step": 80
+    },
+    {
+      "epoch": 0.8723404255319149,
+      "grad_norm": 0.0003809813060797751,
+      "learning_rate": 0.00035425727883076286,
+      "loss": 3.4657,
+      "step": 82
+    },
+    {
+      "epoch": 0.8723404255319149,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 104.9248,
+      "eval_samples_per_second": 2.859,
+      "eval_steps_per_second": 0.048,
+      "step": 82
+    },
+    {
+      "epoch": 0.8936170212765957,
+      "grad_norm": 0.0005144781316630542,
+      "learning_rate": 0.0003507147060424552,
+      "loss": 3.4657,
+      "step": 84
+    },
+    {
+      "epoch": 0.8936170212765957,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 94.2337,
+      "eval_samples_per_second": 3.184,
+      "eval_steps_per_second": 0.053,
+      "step": 84
+    },
+    {
+      "epoch": 0.9148936170212766,
+      "grad_norm": 0.0004984396509826183,
+      "learning_rate": 0.0003471721332541476,
+      "loss": 3.4657,
+      "step": 86
+    },
+    {
+      "epoch": 0.9148936170212766,
+      "eval_loss": 4.103929042816162,
+      "eval_runtime": 93.8643,
+      "eval_samples_per_second": 3.196,
+      "eval_steps_per_second": 0.053,
+      "step": 86
+    },
+    {
+      "epoch": 0.9361702127659575,
+      "grad_norm": 0.00022737662948202342,
+      "learning_rate": 0.00034362956046583994,
+      "loss": 3.4657,
+      "step": 88
+    },
+    {
+      "epoch": 0.9361702127659575,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 135.1229,
+      "eval_samples_per_second": 2.22,
+      "eval_steps_per_second": 0.037,
+      "step": 88
+    },
+    {
+      "epoch": 0.9574468085106383,
+      "grad_norm": 0.0003345514414831996,
+      "learning_rate": 0.00034008698767753234,
+      "loss": 3.4657,
+      "step": 90
+    },
+    {
+      "epoch": 0.9574468085106383,
+      "eval_loss": 4.1039299964904785,
+      "eval_runtime": 141.4817,
+      "eval_samples_per_second": 2.12,
+      "eval_steps_per_second": 0.035,
+      "step": 90
+    },
+    {
+      "epoch": 0.9787234042553191,
+      "grad_norm": 0.00027788232546299696,
+      "learning_rate": 0.0003365444148892247,
+      "loss": 3.4657,
+      "step": 92
+    },
+    {
+      "epoch": 0.9787234042553191,
+      "eval_loss": 4.10392951965332,
+      "eval_runtime": 116.6901,
+      "eval_samples_per_second": 2.571,
+      "eval_steps_per_second": 0.043,
+      "step": 92
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.00033105359761975706,
+      "learning_rate": 0.0003330018421009171,
+      "loss": 3.3006,
+      "step": 94
+    },
+    {
+      "epoch": 1.0,
+      "eval_loss": 4.103929042816162,
+      "eval_runtime": 149.9164,
+      "eval_samples_per_second": 2.001,
+      "eval_steps_per_second": 0.033,
+      "step": 94
+    },
+    {
+      "epoch": 1.0212765957446808,
+      "grad_norm": 0.0003272149770054966,
+      "learning_rate": 0.0003294592693126094,
+      "loss": 3.4657,
+      "step": 96
+    },
+    {
+      "epoch": 1.0212765957446808,
+      "eval_loss": 4.103928565979004,
+      "eval_runtime": 117.8922,
+      "eval_samples_per_second": 2.545,
+      "eval_steps_per_second": 0.042,
+      "step": 96
+    },
+    {
+      "epoch": 1.0425531914893618,
+      "grad_norm": 0.0002339623897569254,
+      "learning_rate": 0.0003259166965243019,
+      "loss": 3.4657,
+      "step": 98
+    },
+    {
+      "epoch": 1.0425531914893618,
+      "eval_loss": 4.10392951965332,
+      "eval_runtime": 94.5849,
+      "eval_samples_per_second": 3.172,
+      "eval_steps_per_second": 0.053,
+      "step": 98
+    },
+    {
+      "epoch": 1.0638297872340425,
+      "grad_norm": 0.0003783780266530812,
+      "learning_rate": 0.0003223741237359942,
+      "loss": 3.4657,
+      "step": 100
+    },
+    {
+      "epoch": 1.0638297872340425,
+      "eval_loss": 4.103929042816162,
+      "eval_runtime": 117.3215,
+      "eval_samples_per_second": 2.557,
+      "eval_steps_per_second": 0.043,
+      "step": 100
+    },
+    {
+      "epoch": 1.0851063829787233,
+      "grad_norm": 0.00021044367167633027,
+      "learning_rate": 0.00031883155094768656,
+      "loss": 3.4657,
+      "step": 102
+    },
+    {
+      "epoch": 1.0851063829787233,
+      "eval_loss": 4.103929042816162,
+      "eval_runtime": 118.2354,
+      "eval_samples_per_second": 2.537,
+      "eval_steps_per_second": 0.042,
+      "step": 102
+    },
+    {
+      "epoch": 1.1063829787234043,
+      "grad_norm": 0.00032288962393067777,
+      "learning_rate": 0.00031528897815937896,
+      "loss": 3.4657,
+      "step": 104
+    },
+    {
+      "epoch": 1.1063829787234043,
+      "eval_loss": 4.103928089141846,
+      "eval_runtime": 111.5984,
+      "eval_samples_per_second": 2.688,
+      "eval_steps_per_second": 0.045,
+      "step": 104
+    },
+    {
+      "epoch": 1.127659574468085,
+      "grad_norm": 0.0002670014800969511,
+      "learning_rate": 0.0003117464053710713,
+      "loss": 3.4657,
+      "step": 106
+    },
+    {
+      "epoch": 1.127659574468085,
+      "eval_loss": 4.103928565979004,
+      "eval_runtime": 121.2363,
+      "eval_samples_per_second": 2.475,
+      "eval_steps_per_second": 0.041,
+      "step": 106
+    },
+    {
+      "epoch": 1.148936170212766,
+      "grad_norm": 0.00021224924421403557,
+      "learning_rate": 0.0003082038325827637,
+      "loss": 3.4657,
+      "step": 108
+    },
+    {
+      "epoch": 1.148936170212766,
+      "eval_loss": 4.103928089141846,
+      "eval_runtime": 102.1198,
+      "eval_samples_per_second": 2.938,
+      "eval_steps_per_second": 0.049,
+      "step": 108
+    },
+    {
+      "epoch": 1.1702127659574468,
+      "grad_norm": 0.0002756421163212508,
+      "learning_rate": 0.00030466125979445604,
+      "loss": 3.4657,
+      "step": 110
+    },
+    {
+      "epoch": 1.1702127659574468,
+      "eval_loss": 4.103928565979004,
+      "eval_runtime": 121.4773,
+      "eval_samples_per_second": 2.47,
+      "eval_steps_per_second": 0.041,
+      "step": 110
+    },
+    {
+      "epoch": 1.1914893617021276,
+      "grad_norm": 0.0002671127731446177,
+      "learning_rate": 0.00030111868700614843,
+      "loss": 3.4657,
+      "step": 112
+    },
+    {
+      "epoch": 1.1914893617021276,
+      "eval_loss": 4.103928565979004,
+      "eval_runtime": 138.1466,
+      "eval_samples_per_second": 2.172,
+      "eval_steps_per_second": 0.036,
+      "step": 112
+    },
+    {
+      "epoch": 1.2127659574468086,
+      "grad_norm": 0.0004021762579213828,
+      "learning_rate": 0.0002975761142178408,
+      "loss": 3.4657,
+      "step": 114
+    },
+    {
+      "epoch": 1.2127659574468086,
+      "eval_loss": 4.103928089141846,
+      "eval_runtime": 101.5548,
+      "eval_samples_per_second": 2.954,
+      "eval_steps_per_second": 0.049,
+      "step": 114
+    },
+    {
+      "epoch": 1.2340425531914894,
+      "grad_norm": 0.0002796368789859116,
+      "learning_rate": 0.0002940335414295331,
+      "loss": 3.4657,
+      "step": 116
+    },
+    {
+      "epoch": 1.2340425531914894,
+      "eval_loss": 4.103928089141846,
+      "eval_runtime": 94.7269,
+      "eval_samples_per_second": 3.167,
+      "eval_steps_per_second": 0.053,
+      "step": 116
+    },
+    {
+      "epoch": 1.2553191489361701,
+      "grad_norm": 0.0003006465267390013,
+      "learning_rate": 0.0002904909686412255,
+      "loss": 3.4657,
+      "step": 118
+    },
+    {
+      "epoch": 1.2553191489361701,
+      "eval_loss": 4.103928089141846,
+      "eval_runtime": 104.4724,
+      "eval_samples_per_second": 2.872,
+      "eval_steps_per_second": 0.048,
+      "step": 118
+    },
+    {
+      "epoch": 1.2765957446808511,
+      "grad_norm": 0.00029228252242319286,
+      "learning_rate": 0.0002869483958529179,
+      "loss": 3.4657,
+      "step": 120
+    },
+    {
+      "epoch": 1.2765957446808511,
+      "eval_loss": 4.103928565979004,
+      "eval_runtime": 108.4499,
+      "eval_samples_per_second": 2.766,
+      "eval_steps_per_second": 0.046,
+      "step": 120
+    },
+    {
+      "epoch": 1.297872340425532,
+      "grad_norm": 0.00032636514515616,
+      "learning_rate": 0.0002834058230646103,
+      "loss": 3.4657,
+      "step": 122
+    },
+    {
+      "epoch": 1.297872340425532,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 94.7255,
+      "eval_samples_per_second": 3.167,
+      "eval_steps_per_second": 0.053,
+      "step": 122
+    },
+    {
+      "epoch": 1.3191489361702127,
+      "grad_norm": 0.00040858419379219413,
+      "learning_rate": 0.00027986325027630265,
+      "loss": 3.4657,
+      "step": 124
+    },
+    {
+      "epoch": 1.3191489361702127,
+      "eval_loss": 4.103928089141846,
+      "eval_runtime": 123.8424,
+      "eval_samples_per_second": 2.422,
+      "eval_steps_per_second": 0.04,
+      "step": 124
+    },
+    {
+      "epoch": 1.3404255319148937,
+      "grad_norm": 0.00040357012767344713,
+      "learning_rate": 0.00027632067748799505,
+      "loss": 3.4657,
+      "step": 126
+    },
+    {
+      "epoch": 1.3404255319148937,
+      "eval_loss": 4.103928089141846,
+      "eval_runtime": 118.9281,
+      "eval_samples_per_second": 2.523,
+      "eval_steps_per_second": 0.042,
+      "step": 126
+    },
+    {
+      "epoch": 1.3617021276595744,
+      "grad_norm": 0.00039930595085024834,
+      "learning_rate": 0.0002727781046996874,
+      "loss": 3.4657,
+      "step": 128
+    },
+    {
+      "epoch": 1.3617021276595744,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 112.1697,
+      "eval_samples_per_second": 2.675,
+      "eval_steps_per_second": 0.045,
+      "step": 128
+    },
+    {
+      "epoch": 1.3829787234042552,
+      "grad_norm": 0.00037575900205411017,
+      "learning_rate": 0.0002692355319113798,
+      "loss": 3.4657,
+      "step": 130
+    },
+    {
+      "epoch": 1.3829787234042552,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 125.21,
+      "eval_samples_per_second": 2.396,
+      "eval_steps_per_second": 0.04,
+      "step": 130
+    },
+    {
+      "epoch": 1.4042553191489362,
+      "grad_norm": 0.00047126636491157115,
+      "learning_rate": 0.00026569295912307213,
+      "loss": 3.4657,
+      "step": 132
+    },
+    {
+      "epoch": 1.4042553191489362,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 135.9147,
+      "eval_samples_per_second": 2.207,
+      "eval_steps_per_second": 0.037,
+      "step": 132
+    },
+    {
+      "epoch": 1.425531914893617,
+      "grad_norm": 0.0005009864689782262,
+      "learning_rate": 0.0002621503863347645,
+      "loss": 3.4657,
+      "step": 134
+    },
+    {
+      "epoch": 1.425531914893617,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 107.4624,
+      "eval_samples_per_second": 2.792,
+      "eval_steps_per_second": 0.047,
+      "step": 134
+    },
+    {
+      "epoch": 1.4468085106382977,
+      "grad_norm": 0.0004678282712120563,
+      "learning_rate": 0.00025860781354645687,
+      "loss": 3.4657,
+      "step": 136
+    },
+    {
+      "epoch": 1.4468085106382977,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 138.7278,
+      "eval_samples_per_second": 2.163,
+      "eval_steps_per_second": 0.036,
+      "step": 136
+    },
+    {
+      "epoch": 1.4680851063829787,
+      "grad_norm": 0.0003415594110265374,
+      "learning_rate": 0.0002550652407581492,
+      "loss": 3.4657,
+      "step": 138
+    },
+    {
+      "epoch": 1.4680851063829787,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 130.873,
+      "eval_samples_per_second": 2.292,
+      "eval_steps_per_second": 0.038,
+      "step": 138
+    },
+    {
+      "epoch": 1.4893617021276595,
+      "grad_norm": 0.00039781673694960773,
+      "learning_rate": 0.0002515226679698416,
+      "loss": 3.4657,
+      "step": 140
+    },
+    {
+      "epoch": 1.4893617021276595,
+      "eval_loss": 4.1039276123046875,
+      "eval_runtime": 127.3509,
+      "eval_samples_per_second": 2.356,
+      "eval_steps_per_second": 0.039,
+      "step": 140
+    },
+    {
+      "epoch": 1.5106382978723403,
+      "grad_norm": 0.0007758048013783991,
+      "learning_rate": 0.000247980095181534,
+      "loss": 3.4657,
+      "step": 142
+    },
+    {
+      "epoch": 1.5106382978723403,
+      "eval_loss": 4.103927135467529,
+      "eval_runtime": 93.1203,
+      "eval_samples_per_second": 3.222,
+      "eval_steps_per_second": 0.054,
+      "step": 142
+    },
+    {
+      "epoch": 1.5319148936170213,
+      "grad_norm": 0.0003962689661420882,
+      "learning_rate": 0.00024443752239322635,
+      "loss": 3.4657,
+      "step": 144
+    },
+    {
+      "epoch": 1.5319148936170213,
+      "eval_loss": 4.103927135467529,
+      "eval_runtime": 92.9702,
+      "eval_samples_per_second": 3.227,
+      "eval_steps_per_second": 0.054,
+      "step": 144
+    },
+    {
+      "epoch": 1.5531914893617023,
+      "grad_norm": 0.00046265136916190386,
+      "learning_rate": 0.00024089494960491872,
+      "loss": 3.4657,
+      "step": 146
+    },
+    {
+      "epoch": 1.5531914893617023,
+      "eval_loss": 4.103926658630371,
+      "eval_runtime": 95.1623,
+      "eval_samples_per_second": 3.153,
+      "eval_steps_per_second": 0.053,
+      "step": 146
+    },
+    {
+      "epoch": 1.574468085106383,
+      "grad_norm": 0.0003199471684638411,
+      "learning_rate": 0.00023735237681661112,
+      "loss": 3.4657,
+      "step": 148
+    },
+    {
+      "epoch": 1.574468085106383,
+      "eval_loss": 4.103927135467529,
+      "eval_runtime": 94.8304,
+      "eval_samples_per_second": 3.164,
+      "eval_steps_per_second": 0.053,
+      "step": 148
+    },
+    {
+      "epoch": 1.5957446808510638,
+      "grad_norm": 0.0009507798822596669,
+      "learning_rate": 0.0002338098040283035,
+      "loss": 3.4657,
+      "step": 150
+    },
+    {
+      "epoch": 1.5957446808510638,
+      "eval_loss": 4.103926658630371,
+      "eval_runtime": 106.393,
+      "eval_samples_per_second": 2.82,
+      "eval_steps_per_second": 0.047,
+      "step": 150
+    },
+    {
+      "epoch": 1.6170212765957448,
+      "grad_norm": 0.0003600665950216353,
+      "learning_rate": 0.00023026723123999586,
+      "loss": 3.4657,
+      "step": 152
+    },
+    {
+      "epoch": 1.6170212765957448,
+      "eval_loss": 4.103927135467529,
+      "eval_runtime": 93.6609,
+      "eval_samples_per_second": 3.203,
+      "eval_steps_per_second": 0.053,
+      "step": 152
+    },
+    {
+      "epoch": 1.6382978723404256,
+      "grad_norm": 0.00046105613000690937,
+      "learning_rate": 0.00022672465845168823,
+      "loss": 3.4657,
+      "step": 154
+    },
+    {
+      "epoch": 1.6382978723404256,
+      "eval_loss": 4.103926658630371,
+      "eval_runtime": 93.7233,
+      "eval_samples_per_second": 3.201,
+      "eval_steps_per_second": 0.053,
+      "step": 154
+    },
+    {
+      "epoch": 1.6595744680851063,
+      "grad_norm": 0.0004165441496297717,
+      "learning_rate": 0.0002231820856633806,
+      "loss": 3.4657,
+      "step": 156
+    },
+    {
+      "epoch": 1.6595744680851063,
+      "eval_loss": 4.103926658630371,
+      "eval_runtime": 93.1558,
+      "eval_samples_per_second": 3.22,
+      "eval_steps_per_second": 0.054,
+      "step": 156
+    },
+    {
+      "epoch": 1.6808510638297873,
+      "grad_norm": 0.00046332846977747977,
+      "learning_rate": 0.00021963951287507297,
+      "loss": 3.4657,
+      "step": 158
+    },
+    {
+      "epoch": 1.6808510638297873,
+      "eval_loss": 4.103925704956055,
+      "eval_runtime": 92.8599,
+      "eval_samples_per_second": 3.231,
+      "eval_steps_per_second": 0.054,
+      "step": 158
+    },
+    {
+      "epoch": 1.702127659574468,
+      "grad_norm": 0.00045260830665938556,
+      "learning_rate": 0.00021609694008676534,
+      "loss": 3.4657,
+      "step": 160
+    },
+    {
+      "epoch": 1.702127659574468,
+      "eval_loss": 4.103925704956055,
+      "eval_runtime": 92.4335,
+      "eval_samples_per_second": 3.246,
+      "eval_steps_per_second": 0.054,
+      "step": 160
+    },
+    {
+      "epoch": 1.7234042553191489,
+      "grad_norm": 0.0006771318148821592,
+      "learning_rate": 0.0002125543672984577,
+      "loss": 3.4657,
+      "step": 162
+    },
+    {
+      "epoch": 1.7234042553191489,
+      "eval_loss": 4.1039252281188965,
+      "eval_runtime": 92.8141,
+      "eval_samples_per_second": 3.232,
+      "eval_steps_per_second": 0.054,
+      "step": 162
+    },
+    {
+      "epoch": 1.7446808510638299,
+      "grad_norm": 0.0002604821929708123,
+      "learning_rate": 0.00020901179451015008,
+      "loss": 3.4657,
+      "step": 164
+    },
+    {
+      "epoch": 1.7446808510638299,
+      "eval_loss": 4.1039252281188965,
+      "eval_runtime": 93.7459,
+      "eval_samples_per_second": 3.2,
+      "eval_steps_per_second": 0.053,
+      "step": 164
+    },
+    {
+      "epoch": 1.7659574468085106,
+      "grad_norm": 0.0006150390254333615,
+      "learning_rate": 0.00020546922172184245,
+      "loss": 3.4657,
+      "step": 166
+    },
+    {
+      "epoch": 1.7659574468085106,
+      "eval_loss": 4.103924751281738,
+      "eval_runtime": 94.135,
+      "eval_samples_per_second": 3.187,
+      "eval_steps_per_second": 0.053,
+      "step": 166
+    },
+    {
+      "epoch": 1.7872340425531914,
+      "grad_norm": 0.0006115163560025394,
+      "learning_rate": 0.00020192664893353482,
+      "loss": 3.4657,
+      "step": 168
+    },
+    {
+      "epoch": 1.7872340425531914,
+      "eval_loss": 4.10392427444458,
+      "eval_runtime": 93.8446,
+      "eval_samples_per_second": 3.197,
+      "eval_steps_per_second": 0.053,
+      "step": 168
+    },
+    {
+      "epoch": 1.8085106382978724,
+      "grad_norm": 0.000593107077293098,
+      "learning_rate": 0.0001983840761452272,
+      "loss": 3.4657,
+      "step": 170
+    },
+    {
+      "epoch": 1.8085106382978724,
+      "eval_loss": 4.10392427444458,
+      "eval_runtime": 92.9188,
+      "eval_samples_per_second": 3.229,
+      "eval_steps_per_second": 0.054,
+      "step": 170
+    },
+    {
+      "epoch": 1.8297872340425532,
+      "grad_norm": 0.0008810166036710143,
+      "learning_rate": 0.00019484150335691958,
+      "loss": 3.4657,
+      "step": 172
+    },
+    {
+      "epoch": 1.8297872340425532,
+      "eval_loss": 4.103923320770264,
+      "eval_runtime": 93.7813,
+      "eval_samples_per_second": 3.199,
+      "eval_steps_per_second": 0.053,
+      "step": 172
+    },
+    {
+      "epoch": 1.851063829787234,
+      "grad_norm": 0.0006031219381839037,
+      "learning_rate": 0.00019129893056861195,
+      "loss": 3.4657,
+      "step": 174
+    },
+    {
+      "epoch": 1.851063829787234,
+      "eval_loss": 4.103923320770264,
+      "eval_runtime": 94.6061,
+      "eval_samples_per_second": 3.171,
+      "eval_steps_per_second": 0.053,
+      "step": 174
+    },
+    {
+      "epoch": 1.872340425531915,
+      "grad_norm": 0.000731271633412689,
+      "learning_rate": 0.0001877563577803043,
+      "loss": 3.4657,
+      "step": 176
+    },
+    {
+      "epoch": 1.872340425531915,
+      "eval_loss": 4.103922367095947,
+      "eval_runtime": 93.6123,
+      "eval_samples_per_second": 3.205,
+      "eval_steps_per_second": 0.053,
+      "step": 176
+    },
+    {
+      "epoch": 1.8936170212765957,
+      "grad_norm": 0.0005839240038767457,
+      "learning_rate": 0.00018421378499199666,
+      "loss": 3.4657,
+      "step": 178
+    },
+    {
+      "epoch": 1.8936170212765957,
+      "eval_loss": 4.103921890258789,
+      "eval_runtime": 93.4167,
+      "eval_samples_per_second": 3.211,
+      "eval_steps_per_second": 0.054,
+      "step": 178
+    },
+    {
+      "epoch": 1.9148936170212765,
+      "grad_norm": 0.0006142465863376856,
+      "learning_rate": 0.00018067121220368906,
+      "loss": 3.4657,
+      "step": 180
+    },
+    {
+      "epoch": 1.9148936170212765,
+      "eval_loss": 4.103921890258789,
+      "eval_runtime": 93.1077,
+      "eval_samples_per_second": 3.222,
+      "eval_steps_per_second": 0.054,
+      "step": 180
+    },
+    {
+      "epoch": 1.9361702127659575,
+      "grad_norm": 0.0005157635896466672,
+      "learning_rate": 0.00017712863941538143,
+      "loss": 3.4657,
+      "step": 182
+    },
+    {
+      "epoch": 1.9361702127659575,
+      "eval_loss": 4.103921890258789,
+      "eval_runtime": 104.9418,
+      "eval_samples_per_second": 2.859,
+      "eval_steps_per_second": 0.048,
+      "step": 182
+    },
+    {
+      "epoch": 1.9574468085106385,
+      "grad_norm": 0.00047715185792185366,
+      "learning_rate": 0.0001735860666270738,
+      "loss": 3.4657,
+      "step": 184
+    },
+    {
+      "epoch": 1.9574468085106385,
+      "eval_loss": 4.103921890258789,
+      "eval_runtime": 143.8492,
+      "eval_samples_per_second": 2.086,
+      "eval_steps_per_second": 0.035,
+      "step": 184
+    },
+    {
+      "epoch": 1.978723404255319,
+      "grad_norm": 0.0006548584206029773,
+      "learning_rate": 0.00017004349383876617,
+      "loss": 3.4657,
+      "step": 186
+    },
+    {
+      "epoch": 1.978723404255319,
+      "eval_loss": 4.103920936584473,
+      "eval_runtime": 93.0081,
+      "eval_samples_per_second": 3.226,
+      "eval_steps_per_second": 0.054,
+      "step": 186
+    },
+    {
+      "epoch": 2.0,
+      "grad_norm": 0.0007490716525353491,
+      "learning_rate": 0.00016650092105045854,
+      "loss": 3.3006,
+      "step": 188
+    },
+    {
+      "epoch": 2.0,
+      "eval_loss": 4.103919982910156,
+      "eval_runtime": 93.8488,
+      "eval_samples_per_second": 3.197,
+      "eval_steps_per_second": 0.053,
+      "step": 188
+    },
+    {
+      "epoch": 2.021276595744681,
+      "grad_norm": 0.001015963382087648,
+      "learning_rate": 0.00016295834826215094,
+      "loss": 3.4657,
+      "step": 190
+    },
+    {
+      "epoch": 2.021276595744681,
+      "eval_loss": 4.103919506072998,
+      "eval_runtime": 93.9004,
+      "eval_samples_per_second": 3.195,
+      "eval_steps_per_second": 0.053,
+      "step": 190
+    },
+    {
+      "epoch": 2.0425531914893615,
+      "grad_norm": 0.004860539920628071,
+      "learning_rate": 0.00015941577547384328,
+      "loss": 3.4657,
+      "step": 192
+    },
+    {
+      "epoch": 2.0425531914893615,
+      "eval_loss": 4.10391902923584,
+      "eval_runtime": 105.0138,
+      "eval_samples_per_second": 2.857,
+      "eval_steps_per_second": 0.048,
+      "step": 192
+    },
+    {
+      "epoch": 2.0638297872340425,
+      "grad_norm": 0.0015178662724792957,
+      "learning_rate": 0.00015587320268553565,
+      "loss": 3.4657,
+      "step": 194
+    },
+    {
+      "epoch": 2.0638297872340425,
+      "eval_loss": 4.103918552398682,
+      "eval_runtime": 94.2786,
+      "eval_samples_per_second": 3.182,
+      "eval_steps_per_second": 0.053,
+      "step": 194
+    },
+    {
+      "epoch": 2.0851063829787235,
+      "grad_norm": 0.0009718618239276111,
+      "learning_rate": 0.00015233062989722802,
+      "loss": 3.4657,
+      "step": 196
+    },
+    {
+      "epoch": 2.0851063829787235,
+      "eval_loss": 4.103917598724365,
+      "eval_runtime": 93.9852,
+      "eval_samples_per_second": 3.192,
+      "eval_steps_per_second": 0.053,
+      "step": 196
+    },
+    {
+      "epoch": 2.106382978723404,
+      "grad_norm": 0.0011399408103898168,
+      "learning_rate": 0.0001487880571089204,
+      "loss": 3.4657,
+      "step": 198
+    },
+    {
+      "epoch": 2.106382978723404,
+      "eval_loss": 4.103916645050049,
+      "eval_runtime": 151.4173,
+      "eval_samples_per_second": 1.981,
+      "eval_steps_per_second": 0.033,
+      "step": 198
+    },
+    {
+      "epoch": 2.127659574468085,
+      "grad_norm": 0.0005256883450783789,
+      "learning_rate": 0.00014524548432061276,
+      "loss": 3.4657,
+      "step": 200
+    },
+    {
+      "epoch": 2.127659574468085,
+      "eval_loss": 4.103916168212891,
+      "eval_runtime": 102.8677,
+      "eval_samples_per_second": 2.916,
+      "eval_steps_per_second": 0.049,
+      "step": 200
+    },
+    {
+      "epoch": 2.148936170212766,
+      "grad_norm": 0.0009113452979363501,
+      "learning_rate": 0.00014170291153230516,
+      "loss": 3.4657,
+      "step": 202
+    },
+    {
+      "epoch": 2.148936170212766,
+      "eval_loss": 4.103915214538574,
+      "eval_runtime": 111.9189,
+      "eval_samples_per_second": 2.681,
+      "eval_steps_per_second": 0.045,
+      "step": 202
+    },
+    {
+      "epoch": 2.1702127659574466,
+      "grad_norm": 0.0010462055215612054,
+      "learning_rate": 0.00013816033874399752,
+      "loss": 3.4657,
+      "step": 204
+    },
+    {
+      "epoch": 2.1702127659574466,
+      "eval_loss": 4.103914260864258,
+      "eval_runtime": 126.7351,
+      "eval_samples_per_second": 2.367,
+      "eval_steps_per_second": 0.039,
+      "step": 204
+    },
+    {
+      "epoch": 2.1914893617021276,
+      "grad_norm": 0.0008617418352514505,
+      "learning_rate": 0.0001346177659556899,
+      "loss": 3.4657,
+      "step": 206
+    },
+    {
+      "epoch": 2.1914893617021276,
+      "eval_loss": 4.1039137840271,
+      "eval_runtime": 124.4636,
+      "eval_samples_per_second": 2.41,
+      "eval_steps_per_second": 0.04,
+      "step": 206
+    },
+    {
+      "epoch": 2.2127659574468086,
+      "grad_norm": 0.0011684081982821226,
+      "learning_rate": 0.00013107519316738224,
+      "loss": 3.4657,
+      "step": 208
+    },
+    {
+      "epoch": 2.2127659574468086,
+      "eval_loss": 4.103911876678467,
+      "eval_runtime": 94.3018,
+      "eval_samples_per_second": 3.181,
+      "eval_steps_per_second": 0.053,
+      "step": 208
+    },
+    {
+      "epoch": 2.2340425531914896,
+      "grad_norm": 0.0008167130872607231,
+      "learning_rate": 0.0001275326203790746,
+      "loss": 3.4657,
+      "step": 210
+    },
+    {
+      "epoch": 2.2340425531914896,
+      "eval_loss": 4.10391092300415,
+      "eval_runtime": 94.2002,
+      "eval_samples_per_second": 3.185,
+      "eval_steps_per_second": 0.053,
+      "step": 210
+    },
+    {
+      "epoch": 2.25531914893617,
+      "grad_norm": 0.0542316697537899,
+      "learning_rate": 0.000123990047590767,
+      "loss": 3.4657,
+      "step": 212
+    },
+    {
+      "epoch": 2.25531914893617,
+      "eval_loss": 4.10391092300415,
+      "eval_runtime": 93.5571,
+      "eval_samples_per_second": 3.207,
+      "eval_steps_per_second": 0.053,
+      "step": 212
+    },
+    {
+      "epoch": 2.276595744680851,
+      "grad_norm": 0.002146972343325615,
+      "learning_rate": 0.00012044747480245936,
+      "loss": 3.4656,
+      "step": 214
+    },
+    {
+      "epoch": 2.276595744680851,
+      "eval_loss": 4.10391092300415,
+      "eval_runtime": 93.6853,
+      "eval_samples_per_second": 3.202,
+      "eval_steps_per_second": 0.053,
+      "step": 214
+    },
+    {
+      "epoch": 2.297872340425532,
+      "grad_norm": 0.0018499704310670495,
+      "learning_rate": 0.00011690490201415174,
+      "loss": 3.4656,
+      "step": 216
+    },
+    {
+      "epoch": 2.297872340425532,
+      "eval_loss": 4.103910446166992,
+      "eval_runtime": 93.9528,
+      "eval_samples_per_second": 3.193,
+      "eval_steps_per_second": 0.053,
+      "step": 216
+    },
+    {
+      "epoch": 2.3191489361702127,
+      "grad_norm": 0.0013631999026983976,
+      "learning_rate": 0.00011336232922584411,
+      "loss": 3.4657,
+      "step": 218
+    },
+    {
+      "epoch": 2.3191489361702127,
+      "eval_loss": 4.103910446166992,
+      "eval_runtime": 93.3718,
+      "eval_samples_per_second": 3.213,
+      "eval_steps_per_second": 0.054,
+      "step": 218
+    },
+    {
+      "epoch": 2.3404255319148937,
+      "grad_norm": 0.0021245332900434732,
+      "learning_rate": 0.00010981975643753648,
+      "loss": 3.4656,
+      "step": 220
+    },
+    {
+      "epoch": 2.3404255319148937,
+      "eval_loss": 4.103908538818359,
+      "eval_runtime": 116.0788,
+      "eval_samples_per_second": 2.584,
+      "eval_steps_per_second": 0.043,
+      "step": 220
+    },
+    {
+      "epoch": 2.3617021276595747,
+      "grad_norm": 0.0015864246524870396,
+      "learning_rate": 0.00010627718364922885,
+      "loss": 3.4657,
+      "step": 222
+    },
+    {
+      "epoch": 2.3617021276595747,
+      "eval_loss": 4.103908061981201,
+      "eval_runtime": 93.8014,
+      "eval_samples_per_second": 3.198,
+      "eval_steps_per_second": 0.053,
+      "step": 222
+    },
+    {
+      "epoch": 2.382978723404255,
+      "grad_norm": 0.0017925110878422856,
+      "learning_rate": 0.00010273461086092122,
+      "loss": 3.4657,
+      "step": 224
+    },
+    {
+      "epoch": 2.382978723404255,
+      "eval_loss": 4.103907108306885,
+      "eval_runtime": 94.17,
+      "eval_samples_per_second": 3.186,
+      "eval_steps_per_second": 0.053,
+      "step": 224
+    },
+    {
+      "epoch": 2.404255319148936,
+      "grad_norm": 0.0022581228986382484,
+      "learning_rate": 9.91920380726136e-05,
+      "loss": 3.4656,
+      "step": 226
+    },
+    {
+      "epoch": 2.404255319148936,
+      "eval_loss": 4.10390567779541,
+      "eval_runtime": 117.7055,
+      "eval_samples_per_second": 2.549,
+      "eval_steps_per_second": 0.042,
+      "step": 226
+    },
+    {
+      "epoch": 2.425531914893617,
+      "grad_norm": 0.0026256099808961153,
+      "learning_rate": 9.564946528430598e-05,
+      "loss": 3.4656,
+      "step": 228
+    },
+    {
+      "epoch": 2.425531914893617,
+      "eval_loss": 4.1039042472839355,
+      "eval_runtime": 128.097,
+      "eval_samples_per_second": 2.342,
+      "eval_steps_per_second": 0.039,
+      "step": 228
+    },
+    {
+      "epoch": 2.4468085106382977,
+      "grad_norm": 0.0022717451211065054,
+      "learning_rate": 9.210689249599833e-05,
+      "loss": 3.4656,
+      "step": 230
+    },
+    {
+      "epoch": 2.4468085106382977,
+      "eval_loss": 4.103902339935303,
+      "eval_runtime": 113.9038,
+      "eval_samples_per_second": 2.634,
+      "eval_steps_per_second": 0.044,
+      "step": 230
+    },
+    {
+      "epoch": 2.4680851063829787,
+      "grad_norm": 0.003249780274927616,
+      "learning_rate": 8.856431970769072e-05,
+      "loss": 3.4656,
+      "step": 232
+    },
+    {
+      "epoch": 2.4680851063829787,
+      "eval_loss": 4.103899955749512,
+      "eval_runtime": 92.3315,
+      "eval_samples_per_second": 3.249,
+      "eval_steps_per_second": 0.054,
+      "step": 232
+    },
+    {
+      "epoch": 2.4893617021276597,
+      "grad_norm": 0.0025626318529248238,
+      "learning_rate": 8.502174691938309e-05,
+      "loss": 3.4656,
+      "step": 234
+    },
+    {
+      "epoch": 2.4893617021276597,
+      "eval_loss": 4.103896617889404,
+      "eval_runtime": 93.3268,
+      "eval_samples_per_second": 3.215,
+      "eval_steps_per_second": 0.054,
+      "step": 234
+    },
+    {
+      "epoch": 2.5106382978723403,
+      "grad_norm": 0.004104019142687321,
+      "learning_rate": 8.147917413107547e-05,
+      "loss": 3.4655,
+      "step": 236
+    },
+    {
+      "epoch": 2.5106382978723403,
+      "eval_loss": 4.103891849517822,
+      "eval_runtime": 93.0172,
+      "eval_samples_per_second": 3.225,
+      "eval_steps_per_second": 0.054,
+      "step": 236
+    },
+    {
+      "epoch": 2.5319148936170213,
+      "grad_norm": 0.005048415157943964,
+      "learning_rate": 7.793660134276782e-05,
+      "loss": 3.4656,
+      "step": 238
+    },
+    {
+      "epoch": 2.5319148936170213,
+      "eval_loss": 4.103887557983398,
+      "eval_runtime": 103.5236,
+      "eval_samples_per_second": 2.898,
+      "eval_steps_per_second": 0.048,
+      "step": 238
+    },
+    {
+      "epoch": 2.5531914893617023,
+      "grad_norm": 0.004755980335175991,
+      "learning_rate": 7.43940285544602e-05,
+      "loss": 3.4656,
+      "step": 240
+    },
+    {
+      "epoch": 2.5531914893617023,
+      "eval_loss": 4.103884220123291,
+      "eval_runtime": 92.2557,
+      "eval_samples_per_second": 3.252,
+      "eval_steps_per_second": 0.054,
+      "step": 240
+    },
+    {
+      "epoch": 2.574468085106383,
+      "grad_norm": 0.00427659647539258,
+      "learning_rate": 7.085145576615258e-05,
+      "loss": 3.4654,
+      "step": 242
+    },
+    {
+      "epoch": 2.574468085106383,
+      "eval_loss": 4.103878498077393,
+      "eval_runtime": 94.0279,
+      "eval_samples_per_second": 3.191,
+      "eval_steps_per_second": 0.053,
+      "step": 242
+    },
+    {
+      "epoch": 2.595744680851064,
+      "grad_norm": 0.0044647688046097755,
+      "learning_rate": 6.730888297784495e-05,
+      "loss": 3.4655,
+      "step": 244
+    },
+    {
+      "epoch": 2.595744680851064,
+      "eval_loss": 4.103872776031494,
+      "eval_runtime": 98.9692,
+      "eval_samples_per_second": 3.031,
+      "eval_steps_per_second": 0.051,
+      "step": 244
+    },
+    {
+      "epoch": 2.617021276595745,
+      "grad_norm": 0.00606452114880085,
+      "learning_rate": 6.37663101895373e-05,
+      "loss": 3.4654,
+      "step": 246
+    },
+    {
+      "epoch": 2.617021276595745,
+      "eval_loss": 4.1038665771484375,
+      "eval_runtime": 130.5821,
+      "eval_samples_per_second": 2.297,
+      "eval_steps_per_second": 0.038,
+      "step": 246
+    },
+    {
+      "epoch": 2.6382978723404253,
+      "grad_norm": 0.0025343315210193396,
+      "learning_rate": 6.022373740122968e-05,
+      "loss": 3.4655,
+      "step": 248
+    },
+    {
+      "epoch": 2.6382978723404253,
+      "eval_loss": 4.103858947753906,
+      "eval_runtime": 110.6408,
+      "eval_samples_per_second": 2.711,
+      "eval_steps_per_second": 0.045,
+      "step": 248
+    },
+    {
+      "epoch": 2.6595744680851063,
+      "grad_norm": 0.004990957211703062,
+      "learning_rate": 5.668116461292206e-05,
+      "loss": 3.4654,
+      "step": 250
+    },
+    {
+      "epoch": 2.6595744680851063,
+      "eval_loss": 4.103850841522217,
+      "eval_runtime": 110.9115,
+      "eval_samples_per_second": 2.705,
+      "eval_steps_per_second": 0.045,
+      "step": 250
+    },
+    {
+      "epoch": 2.6808510638297873,
+      "grad_norm": 0.003452206961810589,
+      "learning_rate": 5.3138591824614426e-05,
+      "loss": 3.4655,
+      "step": 252
+    },
+    {
+      "epoch": 2.6808510638297873,
+      "eval_loss": 4.103842735290527,
+      "eval_runtime": 116.4041,
+      "eval_samples_per_second": 2.577,
+      "eval_steps_per_second": 0.043,
+      "step": 252
+    },
+    {
+      "epoch": 2.702127659574468,
+      "grad_norm": 0.01098131388425827,
+      "learning_rate": 4.95960190363068e-05,
+      "loss": 3.4655,
+      "step": 254
+    },
+    {
+      "epoch": 2.702127659574468,
+      "eval_loss": 4.1038360595703125,
+      "eval_runtime": 120.3016,
+      "eval_samples_per_second": 2.494,
+      "eval_steps_per_second": 0.042,
+      "step": 254
+    },
+    {
+      "epoch": 2.723404255319149,
+      "grad_norm": 0.009479314088821411,
+      "learning_rate": 4.6053446247999166e-05,
+      "loss": 3.4652,
+      "step": 256
+    },
+    {
+      "epoch": 2.723404255319149,
+      "eval_loss": 4.1038289070129395,
+      "eval_runtime": 102.7773,
+      "eval_samples_per_second": 2.919,
+      "eval_steps_per_second": 0.049,
+      "step": 256
+    },
+    {
+      "epoch": 2.74468085106383,
+      "grad_norm": 0.018755914643406868,
+      "learning_rate": 4.251087345969154e-05,
+      "loss": 3.4646,
+      "step": 258
+    },
+    {
+      "epoch": 2.74468085106383,
+      "eval_loss": 4.103816032409668,
+      "eval_runtime": 155.1062,
+      "eval_samples_per_second": 1.934,
+      "eval_steps_per_second": 0.032,
+      "step": 258
+    },
+    {
+      "epoch": 2.7659574468085104,
+      "grad_norm": 0.00831508357077837,
+      "learning_rate": 3.896830067138391e-05,
+      "loss": 3.4654,
+      "step": 260
+    },
+    {
+      "epoch": 2.7659574468085104,
+      "eval_loss": 4.10380220413208,
+      "eval_runtime": 104.9599,
+      "eval_samples_per_second": 2.858,
+      "eval_steps_per_second": 0.048,
+      "step": 260
+    },
+    {
+      "epoch": 2.7872340425531914,
+      "grad_norm": 0.00918419286608696,
+      "learning_rate": 3.542572788307629e-05,
+      "loss": 3.4654,
+      "step": 262
+    },
+    {
+      "epoch": 2.7872340425531914,
+      "eval_loss": 4.103787422180176,
+      "eval_runtime": 121.0806,
+      "eval_samples_per_second": 2.478,
+      "eval_steps_per_second": 0.041,
+      "step": 262
+    },
+    {
+      "epoch": 2.8085106382978724,
+      "grad_norm": 0.010791248641908169,
+      "learning_rate": 3.188315509476865e-05,
+      "loss": 3.4653,
+      "step": 264
+    },
+    {
+      "epoch": 2.8085106382978724,
+      "eval_loss": 4.103773593902588,
+      "eval_runtime": 117.3875,
+      "eval_samples_per_second": 2.556,
+      "eval_steps_per_second": 0.043,
+      "step": 264
+    },
+    {
+      "epoch": 2.829787234042553,
+      "grad_norm": 0.018539341166615486,
+      "learning_rate": 2.834058230646103e-05,
+      "loss": 3.4651,
+      "step": 266
+    },
+    {
+      "epoch": 2.829787234042553,
+      "eval_loss": 4.103758811950684,
+      "eval_runtime": 134.9208,
+      "eval_samples_per_second": 2.224,
+      "eval_steps_per_second": 0.037,
+      "step": 266
+    },
+    {
+      "epoch": 2.851063829787234,
+      "grad_norm": 0.013375967741012573,
+      "learning_rate": 2.47980095181534e-05,
+      "loss": 3.4649,
+      "step": 268
+    },
+    {
+      "epoch": 2.851063829787234,
+      "eval_loss": 4.103740692138672,
+      "eval_runtime": 112.6527,
+      "eval_samples_per_second": 2.663,
+      "eval_steps_per_second": 0.044,
+      "step": 268
+    },
+    {
+      "epoch": 2.872340425531915,
+      "grad_norm": 0.01334660779684782,
+      "learning_rate": 2.125543672984577e-05,
+      "loss": 3.465,
+      "step": 270
+    },
+    {
+      "epoch": 2.872340425531915,
+      "eval_loss": 4.103722095489502,
+      "eval_runtime": 94.5738,
+      "eval_samples_per_second": 3.172,
+      "eval_steps_per_second": 0.053,
+      "step": 270
+    },
+    {
+      "epoch": 2.8936170212765955,
+      "grad_norm": 0.017678698524832726,
+      "learning_rate": 1.7712863941538144e-05,
+      "loss": 3.4647,
+      "step": 272
+    },
+    {
+      "epoch": 2.8936170212765955,
+      "eval_loss": 4.103702068328857,
+      "eval_runtime": 108.8046,
+      "eval_samples_per_second": 2.757,
+      "eval_steps_per_second": 0.046,
+      "step": 272
+    },
+    {
+      "epoch": 2.9148936170212765,
+      "grad_norm": 0.023421209305524826,
+      "learning_rate": 1.4170291153230514e-05,
+      "loss": 3.4642,
+      "step": 274
+    },
+    {
+      "epoch": 2.9148936170212765,
+      "eval_loss": 4.103678226470947,
+      "eval_runtime": 95.2978,
+      "eval_samples_per_second": 3.148,
+      "eval_steps_per_second": 0.052,
+      "step": 274
+    },
+    {
+      "epoch": 2.9361702127659575,
+      "grad_norm": 0.02967756614089012,
+      "learning_rate": 1.0627718364922886e-05,
+      "loss": 3.4649,
+      "step": 276
+    },
+    {
+      "epoch": 2.9361702127659575,
+      "eval_loss": 4.103660583496094,
+      "eval_runtime": 164.156,
+      "eval_samples_per_second": 1.828,
+      "eval_steps_per_second": 0.03,
+      "step": 276
+    },
+    {
+      "epoch": 2.9574468085106385,
+      "grad_norm": 0.04368141293525696,
+      "learning_rate": 7.085145576615257e-06,
+      "loss": 3.4632,
+      "step": 278
+    },
+    {
+      "epoch": 2.9574468085106385,
+      "eval_loss": 4.103642463684082,
+      "eval_runtime": 135.1093,
+      "eval_samples_per_second": 2.22,
+      "eval_steps_per_second": 0.037,
+      "step": 278
+    },
+    {
+      "epoch": 2.978723404255319,
+      "grad_norm": 0.024568898603320122,
+      "learning_rate": 3.5425727883076285e-06,
+      "loss": 3.4644,
+      "step": 280
+    },
+    {
+      "epoch": 2.978723404255319,
+      "eval_loss": 4.103628158569336,
+      "eval_runtime": 111.5348,
+      "eval_samples_per_second": 2.69,
+      "eval_steps_per_second": 0.045,
+      "step": 280
+    },
+    {
+      "epoch": 3.0,
+      "grad_norm": 0.04271033778786659,
+      "learning_rate": 0.0,
+      "loss": 3.2985,
+      "step": 282
+    },
+    {
+      "epoch": 3.0,
+      "eval_loss": 4.103620529174805,
+      "eval_runtime": 122.9717,
+      "eval_samples_per_second": 2.44,
+      "eval_steps_per_second": 0.041,
+      "step": 282
+    }
+  ],
+  "logging_steps": 2,
+  "max_steps": 282,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 523328480700102.0,
+  "train_batch_size": 32,
+  "trial_name": null,
+  "trial_params": {
+    "_wandb": {},
+    "assignments": {},
+    "decay": 0.1,
+    "learning_rate": 0.0004995027631513756,
+    "metric": "eval/loss",
+    "per_device_train_batch_size": 32
+  }
+}

run-bq1dw9ny/checkpoint-282/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d179cbce729053172c2f3e0999512a7fc639544280b821f811c1a6e684dc6c4c
+size 5112

run-yha1u6a1/checkpoint-2/config.json

@@ -0,0 +1,22 @@
+{
+  "_name_or_path": "openai/clip-vit-base-patch32",
+  "architectures": [
+    "CLIPModel"
+  ],
+  "initializer_factor": 1.0,
+  "logit_scale_init_value": 2.6592,
+  "model_type": "clip",
+  "projection_dim": 512,
+  "text_config": {
+    "bos_token_id": 0,
+    "dropout": 0.0,
+    "eos_token_id": 2,
+    "model_type": "clip_text_model"
+  },
+  "torch_dtype": "float32",
+  "transformers_version": "4.42.0.dev0",
+  "vision_config": {
+    "dropout": 0.0,
+    "model_type": "clip_vision_model"
+  }
+}

runs/May23_07-40-12_1eee021f3931/events.out.tfevents.1716472131.1eee021f3931.3634.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7cec984482bd10e65452569acf6b5ea17ac49f20deea438f67de852733e632f0
+size 4184

runs/May26_11-15-36_7c156d14e175/events.out.tfevents.1716742187.7c156d14e175.3825.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:980663982b77393f5ea35bef2b677dfa5b4b68c9b2f48e5a8094545a183efcb1
+size 4184

runs/May27_07-52-26_f4e2ba664464/events.out.tfevents.1716817225.f4e2ba664464.2272.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:91c41b84c166468debb499a314c4686845b6a479a683dc107086b103f9882783
+size 4184

runs/May29_14-16-19_00a42a5d637b/events.out.tfevents.1716992396.00a42a5d637b.22663.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5bc8ad1f4e57d3a25f471d1aaf6c186d01f953992d26a1e3b661d042a1a94c65
+size 4184

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:10e3da5a2c2894297cb56e6aa655f2850fc6a872b1aff8030670c382688d25ff
+oid sha256:d179cbce729053172c2f3e0999512a7fc639544280b821f811c1a6e684dc6c4c
 size 5112