Upload folder using huggingface_hub

README.md

@@ -0,0 +1,28 @@
+
+---
+tags:
+- autotrain
+- text-regression
+base_model: distilbert/distilbert-base-uncased
+widget:
+- text: "I love AutoTrain"
+datasets:
+- samtuckervegan/relevancy
+---
+
+# Model Trained Using AutoTrain
+
+- Problem type: Text Regression
+
+## Validation Metrics
+loss: 0.04224373400211334
+
+mse: 0.04224866256117821
+
+mae: 0.15929067134857178
+
+r2: 0.23554444313049316
+
+rmse: 0.20554479453680702
+
+explained_variance: 0.23569148778915405

checkpoint-7980/config.json

@@ -0,0 +1,32 @@
+{
+  "_name_or_path": "distilbert/distilbert-base-uncased",
+  "_num_labels": 1,
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "id2label": {
+    "0": "target"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "target": 0
+  },
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "regression",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.0",
+  "vocab_size": 30522
+}

checkpoint-7980/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7eebdfe40823a90e15ce6d5701dbe3da53f026b8ce64e6c583ed64be7aa1150d
+size 267829484

checkpoint-7980/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40a7f4ac42665a4375260fc08a68309354b66af0d54c9f0e13c48d8c620402da
+size 535721146

checkpoint-7980/rng_state_0.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:191b7d80c7eefd89ba0c9e105e927a40fadf120d2bb3de1043b025b16923c222
+size 15024

checkpoint-7980/rng_state_1.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:265d4e8794d69f71649f53e597bae382154bb6b8016836972c3c5b93e2842fa7
+size 15024

checkpoint-7980/rng_state_2.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b915cb9898748bc446c463bdd075e28a9a83539067b921a4decdea6a83d093d9
+size 15024

checkpoint-7980/rng_state_3.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cf64ec88b98d7b2fb48824a98c28789d1de90d7fa5bdda84a74e8044466972e
+size 15024

checkpoint-7980/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a0c3b48e85066c0f06c8412dd982a0730376407d87a004332409cdaabe7fa80b
+size 1064

checkpoint-7980/trainer_state.json

@@ -0,0 +1,2314 @@
+{
+  "best_metric": 0.04224373400211334,
+  "best_model_checkpoint": "relevancy/checkpoint-7980",
+  "epoch": 3.0,
+  "eval_steps": 500,
+  "global_step": 7980,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.009398496240601503,
+      "grad_norm": 7.968059539794922,
+      "learning_rate": 1.5037593984962406e-06,
+      "loss": 0.5421,
+      "step": 25
+    },
+    {
+      "epoch": 0.018796992481203006,
+      "grad_norm": 3.0120198726654053,
+      "learning_rate": 3.070175438596491e-06,
+      "loss": 0.3055,
+      "step": 50
+    },
+    {
+      "epoch": 0.02819548872180451,
+      "grad_norm": 0.7311064600944519,
+      "learning_rate": 4.636591478696742e-06,
+      "loss": 0.0816,
+      "step": 75
+    },
+    {
+      "epoch": 0.03759398496240601,
+      "grad_norm": 1.2286335229873657,
+      "learning_rate": 6.203007518796992e-06,
+      "loss": 0.0617,
+      "step": 100
+    },
+    {
+      "epoch": 0.046992481203007516,
+      "grad_norm": 0.5518651008605957,
+      "learning_rate": 7.769423558897243e-06,
+      "loss": 0.0552,
+      "step": 125
+    },
+    {
+      "epoch": 0.05639097744360902,
+      "grad_norm": 0.5839918255805969,
+      "learning_rate": 9.335839598997493e-06,
+      "loss": 0.062,
+      "step": 150
+    },
+    {
+      "epoch": 0.06578947368421052,
+      "grad_norm": 0.6332781314849854,
+      "learning_rate": 1.0902255639097744e-05,
+      "loss": 0.0528,
+      "step": 175
+    },
+    {
+      "epoch": 0.07518796992481203,
+      "grad_norm": 0.4804973304271698,
+      "learning_rate": 1.2468671679197996e-05,
+      "loss": 0.0563,
+      "step": 200
+    },
+    {
+      "epoch": 0.08458646616541353,
+      "grad_norm": 0.4023095667362213,
+      "learning_rate": 1.4035087719298246e-05,
+      "loss": 0.0607,
+      "step": 225
+    },
+    {
+      "epoch": 0.09398496240601503,
+      "grad_norm": 1.1123720407485962,
+      "learning_rate": 1.5601503759398496e-05,
+      "loss": 0.0591,
+      "step": 250
+    },
+    {
+      "epoch": 0.10338345864661654,
+      "grad_norm": 0.9853416085243225,
+      "learning_rate": 1.7167919799498746e-05,
+      "loss": 0.0547,
+      "step": 275
+    },
+    {
+      "epoch": 0.11278195488721804,
+      "grad_norm": 1.4674689769744873,
+      "learning_rate": 1.8734335839599e-05,
+      "loss": 0.0554,
+      "step": 300
+    },
+    {
+      "epoch": 0.12218045112781954,
+      "grad_norm": 1.15469229221344,
+      "learning_rate": 2.0300751879699247e-05,
+      "loss": 0.0577,
+      "step": 325
+    },
+    {
+      "epoch": 0.13157894736842105,
+      "grad_norm": 0.7320540547370911,
+      "learning_rate": 2.18671679197995e-05,
+      "loss": 0.0562,
+      "step": 350
+    },
+    {
+      "epoch": 0.14097744360902256,
+      "grad_norm": 0.3801420032978058,
+      "learning_rate": 2.3433583959899748e-05,
+      "loss": 0.0533,
+      "step": 375
+    },
+    {
+      "epoch": 0.15037593984962405,
+      "grad_norm": 0.9615464806556702,
+      "learning_rate": 2.5e-05,
+      "loss": 0.0589,
+      "step": 400
+    },
+    {
+      "epoch": 0.15977443609022557,
+      "grad_norm": 0.3856410086154938,
+      "learning_rate": 2.656641604010025e-05,
+      "loss": 0.0548,
+      "step": 425
+    },
+    {
+      "epoch": 0.16917293233082706,
+      "grad_norm": 1.2804025411605835,
+      "learning_rate": 2.8132832080200505e-05,
+      "loss": 0.0561,
+      "step": 450
+    },
+    {
+      "epoch": 0.17857142857142858,
+      "grad_norm": 0.9871304035186768,
+      "learning_rate": 2.9699248120300755e-05,
+      "loss": 0.052,
+      "step": 475
+    },
+    {
+      "epoch": 0.18796992481203006,
+      "grad_norm": 1.0679211616516113,
+      "learning_rate": 3.1265664160401006e-05,
+      "loss": 0.0519,
+      "step": 500
+    },
+    {
+      "epoch": 0.19736842105263158,
+      "grad_norm": 0.2823966145515442,
+      "learning_rate": 3.2832080200501256e-05,
+      "loss": 0.053,
+      "step": 525
+    },
+    {
+      "epoch": 0.20676691729323307,
+      "grad_norm": 1.1498183012008667,
+      "learning_rate": 3.4398496240601506e-05,
+      "loss": 0.054,
+      "step": 550
+    },
+    {
+      "epoch": 0.2161654135338346,
+      "grad_norm": 1.0924912691116333,
+      "learning_rate": 3.5964912280701756e-05,
+      "loss": 0.0503,
+      "step": 575
+    },
+    {
+      "epoch": 0.22556390977443608,
+      "grad_norm": 1.0172274112701416,
+      "learning_rate": 3.753132832080201e-05,
+      "loss": 0.0544,
+      "step": 600
+    },
+    {
+      "epoch": 0.2349624060150376,
+      "grad_norm": 0.4584837853908539,
+      "learning_rate": 3.909774436090226e-05,
+      "loss": 0.0536,
+      "step": 625
+    },
+    {
+      "epoch": 0.24436090225563908,
+      "grad_norm": 0.38874387741088867,
+      "learning_rate": 4.066416040100251e-05,
+      "loss": 0.0535,
+      "step": 650
+    },
+    {
+      "epoch": 0.25375939849624063,
+      "grad_norm": 1.104377031326294,
+      "learning_rate": 4.223057644110276e-05,
+      "loss": 0.0495,
+      "step": 675
+    },
+    {
+      "epoch": 0.2631578947368421,
+      "grad_norm": 0.40505892038345337,
+      "learning_rate": 4.379699248120301e-05,
+      "loss": 0.0539,
+      "step": 700
+    },
+    {
+      "epoch": 0.2725563909774436,
+      "grad_norm": 0.6633829474449158,
+      "learning_rate": 4.536340852130326e-05,
+      "loss": 0.0533,
+      "step": 725
+    },
+    {
+      "epoch": 0.2819548872180451,
+      "grad_norm": 0.49940770864486694,
+      "learning_rate": 4.6929824561403515e-05,
+      "loss": 0.0538,
+      "step": 750
+    },
+    {
+      "epoch": 0.29135338345864664,
+      "grad_norm": 0.9451118111610413,
+      "learning_rate": 4.849624060150376e-05,
+      "loss": 0.0509,
+      "step": 775
+    },
+    {
+      "epoch": 0.3007518796992481,
+      "grad_norm": 0.47112253308296204,
+      "learning_rate": 4.9993038150932894e-05,
+      "loss": 0.058,
+      "step": 800
+    },
+    {
+      "epoch": 0.3101503759398496,
+      "grad_norm": 0.6366233229637146,
+      "learning_rate": 4.9818991924255085e-05,
+      "loss": 0.0509,
+      "step": 825
+    },
+    {
+      "epoch": 0.31954887218045114,
+      "grad_norm": 1.5148544311523438,
+      "learning_rate": 4.9644945697577276e-05,
+      "loss": 0.0554,
+      "step": 850
+    },
+    {
+      "epoch": 0.32894736842105265,
+      "grad_norm": 0.8564983606338501,
+      "learning_rate": 4.9470899470899475e-05,
+      "loss": 0.052,
+      "step": 875
+    },
+    {
+      "epoch": 0.3383458646616541,
+      "grad_norm": 0.32261064648628235,
+      "learning_rate": 4.9296853244221666e-05,
+      "loss": 0.0518,
+      "step": 900
+    },
+    {
+      "epoch": 0.34774436090225563,
+      "grad_norm": 0.3869636356830597,
+      "learning_rate": 4.912280701754386e-05,
+      "loss": 0.054,
+      "step": 925
+    },
+    {
+      "epoch": 0.35714285714285715,
+      "grad_norm": 0.5892410278320312,
+      "learning_rate": 4.8948760790866056e-05,
+      "loss": 0.0542,
+      "step": 950
+    },
+    {
+      "epoch": 0.36654135338345867,
+      "grad_norm": 0.48280876874923706,
+      "learning_rate": 4.8774714564188254e-05,
+      "loss": 0.0525,
+      "step": 975
+    },
+    {
+      "epoch": 0.37593984962406013,
+      "grad_norm": 0.3844253718852997,
+      "learning_rate": 4.8600668337510445e-05,
+      "loss": 0.0522,
+      "step": 1000
+    },
+    {
+      "epoch": 0.38533834586466165,
+      "grad_norm": 0.2661376893520355,
+      "learning_rate": 4.8426622110832637e-05,
+      "loss": 0.0495,
+      "step": 1025
+    },
+    {
+      "epoch": 0.39473684210526316,
+      "grad_norm": 0.44888997077941895,
+      "learning_rate": 4.8252575884154835e-05,
+      "loss": 0.0522,
+      "step": 1050
+    },
+    {
+      "epoch": 0.4041353383458647,
+      "grad_norm": 0.939777672290802,
+      "learning_rate": 4.8078529657477026e-05,
+      "loss": 0.054,
+      "step": 1075
+    },
+    {
+      "epoch": 0.41353383458646614,
+      "grad_norm": 0.4625527560710907,
+      "learning_rate": 4.7904483430799224e-05,
+      "loss": 0.0474,
+      "step": 1100
+    },
+    {
+      "epoch": 0.42293233082706766,
+      "grad_norm": 0.8745533227920532,
+      "learning_rate": 4.7730437204121416e-05,
+      "loss": 0.0489,
+      "step": 1125
+    },
+    {
+      "epoch": 0.4323308270676692,
+      "grad_norm": 0.48227739334106445,
+      "learning_rate": 4.755639097744361e-05,
+      "loss": 0.0541,
+      "step": 1150
+    },
+    {
+      "epoch": 0.4417293233082707,
+      "grad_norm": 0.7436401844024658,
+      "learning_rate": 4.7382344750765805e-05,
+      "loss": 0.0503,
+      "step": 1175
+    },
+    {
+      "epoch": 0.45112781954887216,
+      "grad_norm": 0.594455897808075,
+      "learning_rate": 4.7208298524088003e-05,
+      "loss": 0.0477,
+      "step": 1200
+    },
+    {
+      "epoch": 0.4605263157894737,
+      "grad_norm": 0.27695509791374207,
+      "learning_rate": 4.7034252297410195e-05,
+      "loss": 0.0503,
+      "step": 1225
+    },
+    {
+      "epoch": 0.4699248120300752,
+      "grad_norm": 0.2683180272579193,
+      "learning_rate": 4.6860206070732386e-05,
+      "loss": 0.0532,
+      "step": 1250
+    },
+    {
+      "epoch": 0.4793233082706767,
+      "grad_norm": 0.7848153114318848,
+      "learning_rate": 4.6686159844054584e-05,
+      "loss": 0.0465,
+      "step": 1275
+    },
+    {
+      "epoch": 0.48872180451127817,
+      "grad_norm": 0.5359299778938293,
+      "learning_rate": 4.6512113617376776e-05,
+      "loss": 0.0483,
+      "step": 1300
+    },
+    {
+      "epoch": 0.4981203007518797,
+      "grad_norm": 0.439117968082428,
+      "learning_rate": 4.633806739069897e-05,
+      "loss": 0.0513,
+      "step": 1325
+    },
+    {
+      "epoch": 0.5075187969924813,
+      "grad_norm": 0.3284767270088196,
+      "learning_rate": 4.6164021164021165e-05,
+      "loss": 0.0477,
+      "step": 1350
+    },
+    {
+      "epoch": 0.5169172932330827,
+      "grad_norm": 0.542739748954773,
+      "learning_rate": 4.5989974937343364e-05,
+      "loss": 0.0503,
+      "step": 1375
+    },
+    {
+      "epoch": 0.5263157894736842,
+      "grad_norm": 0.6036680936813354,
+      "learning_rate": 4.5815928710665555e-05,
+      "loss": 0.0495,
+      "step": 1400
+    },
+    {
+      "epoch": 0.5357142857142857,
+      "grad_norm": 0.37653249502182007,
+      "learning_rate": 4.5641882483987746e-05,
+      "loss": 0.0486,
+      "step": 1425
+    },
+    {
+      "epoch": 0.5451127819548872,
+      "grad_norm": 0.35741862654685974,
+      "learning_rate": 4.5467836257309945e-05,
+      "loss": 0.0469,
+      "step": 1450
+    },
+    {
+      "epoch": 0.5545112781954887,
+      "grad_norm": 0.3325374126434326,
+      "learning_rate": 4.5293790030632136e-05,
+      "loss": 0.0451,
+      "step": 1475
+    },
+    {
+      "epoch": 0.5639097744360902,
+      "grad_norm": 0.5224691033363342,
+      "learning_rate": 4.5119743803954334e-05,
+      "loss": 0.0488,
+      "step": 1500
+    },
+    {
+      "epoch": 0.5733082706766918,
+      "grad_norm": 0.8083314299583435,
+      "learning_rate": 4.4945697577276526e-05,
+      "loss": 0.0486,
+      "step": 1525
+    },
+    {
+      "epoch": 0.5827067669172933,
+      "grad_norm": 0.7639009952545166,
+      "learning_rate": 4.477165135059872e-05,
+      "loss": 0.0438,
+      "step": 1550
+    },
+    {
+      "epoch": 0.5921052631578947,
+      "grad_norm": 0.5051703453063965,
+      "learning_rate": 4.4597605123920915e-05,
+      "loss": 0.0477,
+      "step": 1575
+    },
+    {
+      "epoch": 0.6015037593984962,
+      "grad_norm": 0.901390552520752,
+      "learning_rate": 4.442355889724311e-05,
+      "loss": 0.045,
+      "step": 1600
+    },
+    {
+      "epoch": 0.6109022556390977,
+      "grad_norm": 0.32415810227394104,
+      "learning_rate": 4.4249512670565305e-05,
+      "loss": 0.0437,
+      "step": 1625
+    },
+    {
+      "epoch": 0.6203007518796992,
+      "grad_norm": 0.4060637652873993,
+      "learning_rate": 4.4075466443887496e-05,
+      "loss": 0.0467,
+      "step": 1650
+    },
+    {
+      "epoch": 0.6296992481203008,
+      "grad_norm": 0.29624292254447937,
+      "learning_rate": 4.3901420217209694e-05,
+      "loss": 0.0491,
+      "step": 1675
+    },
+    {
+      "epoch": 0.6390977443609023,
+      "grad_norm": 0.2997782826423645,
+      "learning_rate": 4.3727373990531886e-05,
+      "loss": 0.049,
+      "step": 1700
+    },
+    {
+      "epoch": 0.6484962406015038,
+      "grad_norm": 0.8522358536720276,
+      "learning_rate": 4.355332776385408e-05,
+      "loss": 0.0461,
+      "step": 1725
+    },
+    {
+      "epoch": 0.6578947368421053,
+      "grad_norm": 0.6091852784156799,
+      "learning_rate": 4.3379281537176275e-05,
+      "loss": 0.0497,
+      "step": 1750
+    },
+    {
+      "epoch": 0.6672932330827067,
+      "grad_norm": 0.3918594419956207,
+      "learning_rate": 4.3205235310498473e-05,
+      "loss": 0.051,
+      "step": 1775
+    },
+    {
+      "epoch": 0.6766917293233082,
+      "grad_norm": 0.44399502873420715,
+      "learning_rate": 4.3031189083820665e-05,
+      "loss": 0.0452,
+      "step": 1800
+    },
+    {
+      "epoch": 0.6860902255639098,
+      "grad_norm": 0.4616807699203491,
+      "learning_rate": 4.2857142857142856e-05,
+      "loss": 0.0466,
+      "step": 1825
+    },
+    {
+      "epoch": 0.6954887218045113,
+      "grad_norm": 0.3829876780509949,
+      "learning_rate": 4.2683096630465054e-05,
+      "loss": 0.0458,
+      "step": 1850
+    },
+    {
+      "epoch": 0.7048872180451128,
+      "grad_norm": 1.2150853872299194,
+      "learning_rate": 4.2509050403787246e-05,
+      "loss": 0.044,
+      "step": 1875
+    },
+    {
+      "epoch": 0.7142857142857143,
+      "grad_norm": 0.3793696165084839,
+      "learning_rate": 4.2335004177109444e-05,
+      "loss": 0.0464,
+      "step": 1900
+    },
+    {
+      "epoch": 0.7236842105263158,
+      "grad_norm": 0.3776131272315979,
+      "learning_rate": 4.2160957950431635e-05,
+      "loss": 0.0465,
+      "step": 1925
+    },
+    {
+      "epoch": 0.7330827067669173,
+      "grad_norm": 0.47644665837287903,
+      "learning_rate": 4.198691172375383e-05,
+      "loss": 0.0477,
+      "step": 1950
+    },
+    {
+      "epoch": 0.7424812030075187,
+      "grad_norm": 0.17218613624572754,
+      "learning_rate": 4.1812865497076025e-05,
+      "loss": 0.0437,
+      "step": 1975
+    },
+    {
+      "epoch": 0.7518796992481203,
+      "grad_norm": 0.5325891971588135,
+      "learning_rate": 4.163881927039822e-05,
+      "loss": 0.049,
+      "step": 2000
+    },
+    {
+      "epoch": 0.7612781954887218,
+      "grad_norm": 0.3780176341533661,
+      "learning_rate": 4.1464773043720415e-05,
+      "loss": 0.0472,
+      "step": 2025
+    },
+    {
+      "epoch": 0.7706766917293233,
+      "grad_norm": 0.5997790694236755,
+      "learning_rate": 4.1290726817042606e-05,
+      "loss": 0.0492,
+      "step": 2050
+    },
+    {
+      "epoch": 0.7800751879699248,
+      "grad_norm": 0.572754979133606,
+      "learning_rate": 4.1116680590364804e-05,
+      "loss": 0.0488,
+      "step": 2075
+    },
+    {
+      "epoch": 0.7894736842105263,
+      "grad_norm": 0.5054708123207092,
+      "learning_rate": 4.0942634363687e-05,
+      "loss": 0.0447,
+      "step": 2100
+    },
+    {
+      "epoch": 0.7988721804511278,
+      "grad_norm": 0.3036394417285919,
+      "learning_rate": 4.076858813700919e-05,
+      "loss": 0.044,
+      "step": 2125
+    },
+    {
+      "epoch": 0.8082706766917294,
+      "grad_norm": 0.2748730480670929,
+      "learning_rate": 4.0594541910331385e-05,
+      "loss": 0.0478,
+      "step": 2150
+    },
+    {
+      "epoch": 0.8176691729323309,
+      "grad_norm": 0.23642213642597198,
+      "learning_rate": 4.042049568365358e-05,
+      "loss": 0.0458,
+      "step": 2175
+    },
+    {
+      "epoch": 0.8270676691729323,
+      "grad_norm": 0.2958623766899109,
+      "learning_rate": 4.0246449456975775e-05,
+      "loss": 0.0428,
+      "step": 2200
+    },
+    {
+      "epoch": 0.8364661654135338,
+      "grad_norm": 1.332788109779358,
+      "learning_rate": 4.0072403230297966e-05,
+      "loss": 0.0481,
+      "step": 2225
+    },
+    {
+      "epoch": 0.8458646616541353,
+      "grad_norm": 0.3045213222503662,
+      "learning_rate": 3.9898357003620164e-05,
+      "loss": 0.0445,
+      "step": 2250
+    },
+    {
+      "epoch": 0.8552631578947368,
+      "grad_norm": 0.2599555253982544,
+      "learning_rate": 3.9724310776942356e-05,
+      "loss": 0.0442,
+      "step": 2275
+    },
+    {
+      "epoch": 0.8646616541353384,
+      "grad_norm": 0.3255714178085327,
+      "learning_rate": 3.9550264550264554e-05,
+      "loss": 0.0483,
+      "step": 2300
+    },
+    {
+      "epoch": 0.8740601503759399,
+      "grad_norm": 1.2524524927139282,
+      "learning_rate": 3.9376218323586745e-05,
+      "loss": 0.0426,
+      "step": 2325
+    },
+    {
+      "epoch": 0.8834586466165414,
+      "grad_norm": 0.2243344783782959,
+      "learning_rate": 3.920217209690894e-05,
+      "loss": 0.0424,
+      "step": 2350
+    },
+    {
+      "epoch": 0.8928571428571429,
+      "grad_norm": 0.3369060158729553,
+      "learning_rate": 3.9028125870231135e-05,
+      "loss": 0.0477,
+      "step": 2375
+    },
+    {
+      "epoch": 0.9022556390977443,
+      "grad_norm": 0.2334790974855423,
+      "learning_rate": 3.885407964355333e-05,
+      "loss": 0.0425,
+      "step": 2400
+    },
+    {
+      "epoch": 0.9116541353383458,
+      "grad_norm": 0.5213820338249207,
+      "learning_rate": 3.8680033416875525e-05,
+      "loss": 0.0472,
+      "step": 2425
+    },
+    {
+      "epoch": 0.9210526315789473,
+      "grad_norm": 0.31274598836898804,
+      "learning_rate": 3.8505987190197716e-05,
+      "loss": 0.045,
+      "step": 2450
+    },
+    {
+      "epoch": 0.9304511278195489,
+      "grad_norm": 0.36253780126571655,
+      "learning_rate": 3.8331940963519914e-05,
+      "loss": 0.0442,
+      "step": 2475
+    },
+    {
+      "epoch": 0.9398496240601504,
+      "grad_norm": 0.6255716681480408,
+      "learning_rate": 3.815789473684211e-05,
+      "loss": 0.0428,
+      "step": 2500
+    },
+    {
+      "epoch": 0.9492481203007519,
+      "grad_norm": 0.3290731906890869,
+      "learning_rate": 3.79838485101643e-05,
+      "loss": 0.0462,
+      "step": 2525
+    },
+    {
+      "epoch": 0.9586466165413534,
+      "grad_norm": 0.747561514377594,
+      "learning_rate": 3.7809802283486495e-05,
+      "loss": 0.0505,
+      "step": 2550
+    },
+    {
+      "epoch": 0.9680451127819549,
+      "grad_norm": 0.29496845602989197,
+      "learning_rate": 3.763575605680869e-05,
+      "loss": 0.046,
+      "step": 2575
+    },
+    {
+      "epoch": 0.9774436090225563,
+      "grad_norm": 0.23728559911251068,
+      "learning_rate": 3.7461709830130885e-05,
+      "loss": 0.051,
+      "step": 2600
+    },
+    {
+      "epoch": 0.9868421052631579,
+      "grad_norm": 0.39210203289985657,
+      "learning_rate": 3.7287663603453076e-05,
+      "loss": 0.0456,
+      "step": 2625
+    },
+    {
+      "epoch": 0.9962406015037594,
+      "grad_norm": 0.4230247139930725,
+      "learning_rate": 3.7113617376775274e-05,
+      "loss": 0.0479,
+      "step": 2650
+    },
+    {
+      "epoch": 1.0,
+      "eval_explained_variance": 0.19852977991104126,
+      "eval_loss": 0.044305671006441116,
+      "eval_mae": 0.16591642796993256,
+      "eval_mse": 0.0443120077252388,
+      "eval_r2": 0.19820988178253174,
+      "eval_rmse": 0.21050417507792762,
+      "eval_runtime": 5.0465,
+      "eval_samples_per_second": 4216.201,
+      "eval_steps_per_second": 65.987,
+      "step": 2660
+    },
+    {
+      "epoch": 1.005639097744361,
+      "grad_norm": 0.6470463871955872,
+      "learning_rate": 3.6939571150097466e-05,
+      "loss": 0.0435,
+      "step": 2675
+    },
+    {
+      "epoch": 1.0150375939849625,
+      "grad_norm": 0.37490159273147583,
+      "learning_rate": 3.6765524923419664e-05,
+      "loss": 0.0389,
+      "step": 2700
+    },
+    {
+      "epoch": 1.0244360902255638,
+      "grad_norm": 0.42144620418548584,
+      "learning_rate": 3.6591478696741855e-05,
+      "loss": 0.0388,
+      "step": 2725
+    },
+    {
+      "epoch": 1.0338345864661653,
+      "grad_norm": 0.384360671043396,
+      "learning_rate": 3.6417432470064053e-05,
+      "loss": 0.0463,
+      "step": 2750
+    },
+    {
+      "epoch": 1.0432330827067668,
+      "grad_norm": 0.8169638514518738,
+      "learning_rate": 3.6243386243386245e-05,
+      "loss": 0.0439,
+      "step": 2775
+    },
+    {
+      "epoch": 1.0526315789473684,
+      "grad_norm": 0.7266212105751038,
+      "learning_rate": 3.606934001670844e-05,
+      "loss": 0.0439,
+      "step": 2800
+    },
+    {
+      "epoch": 1.0620300751879699,
+      "grad_norm": 0.5463135838508606,
+      "learning_rate": 3.5895293790030634e-05,
+      "loss": 0.0473,
+      "step": 2825
+    },
+    {
+      "epoch": 1.0714285714285714,
+      "grad_norm": 0.39723679423332214,
+      "learning_rate": 3.5721247563352826e-05,
+      "loss": 0.0403,
+      "step": 2850
+    },
+    {
+      "epoch": 1.080827067669173,
+      "grad_norm": 0.8675305843353271,
+      "learning_rate": 3.5547201336675024e-05,
+      "loss": 0.0454,
+      "step": 2875
+    },
+    {
+      "epoch": 1.0902255639097744,
+      "grad_norm": 0.27421316504478455,
+      "learning_rate": 3.537315510999722e-05,
+      "loss": 0.0428,
+      "step": 2900
+    },
+    {
+      "epoch": 1.099624060150376,
+      "grad_norm": 0.27585676312446594,
+      "learning_rate": 3.519910888331941e-05,
+      "loss": 0.043,
+      "step": 2925
+    },
+    {
+      "epoch": 1.1090225563909775,
+      "grad_norm": 0.7620128393173218,
+      "learning_rate": 3.5025062656641605e-05,
+      "loss": 0.0435,
+      "step": 2950
+    },
+    {
+      "epoch": 1.118421052631579,
+      "grad_norm": 0.5449320077896118,
+      "learning_rate": 3.48510164299638e-05,
+      "loss": 0.0424,
+      "step": 2975
+    },
+    {
+      "epoch": 1.1278195488721805,
+      "grad_norm": 0.3466412425041199,
+      "learning_rate": 3.4676970203285995e-05,
+      "loss": 0.0397,
+      "step": 3000
+    },
+    {
+      "epoch": 1.137218045112782,
+      "grad_norm": 0.2692334055900574,
+      "learning_rate": 3.4502923976608186e-05,
+      "loss": 0.0433,
+      "step": 3025
+    },
+    {
+      "epoch": 1.1466165413533835,
+      "grad_norm": 0.3983423411846161,
+      "learning_rate": 3.4328877749930384e-05,
+      "loss": 0.0454,
+      "step": 3050
+    },
+    {
+      "epoch": 1.156015037593985,
+      "grad_norm": 0.28615260124206543,
+      "learning_rate": 3.4154831523252576e-05,
+      "loss": 0.0458,
+      "step": 3075
+    },
+    {
+      "epoch": 1.1654135338345863,
+      "grad_norm": 0.6662613153457642,
+      "learning_rate": 3.3980785296574774e-05,
+      "loss": 0.0386,
+      "step": 3100
+    },
+    {
+      "epoch": 1.1748120300751879,
+      "grad_norm": 0.7193450331687927,
+      "learning_rate": 3.3806739069896965e-05,
+      "loss": 0.0465,
+      "step": 3125
+    },
+    {
+      "epoch": 1.1842105263157894,
+      "grad_norm": 0.6499005556106567,
+      "learning_rate": 3.363269284321916e-05,
+      "loss": 0.0473,
+      "step": 3150
+    },
+    {
+      "epoch": 1.193609022556391,
+      "grad_norm": 0.45330649614334106,
+      "learning_rate": 3.3458646616541355e-05,
+      "loss": 0.0413,
+      "step": 3175
+    },
+    {
+      "epoch": 1.2030075187969924,
+      "grad_norm": 0.2722605764865875,
+      "learning_rate": 3.328460038986355e-05,
+      "loss": 0.0407,
+      "step": 3200
+    },
+    {
+      "epoch": 1.212406015037594,
+      "grad_norm": 0.3249507546424866,
+      "learning_rate": 3.3110554163185744e-05,
+      "loss": 0.0431,
+      "step": 3225
+    },
+    {
+      "epoch": 1.2218045112781954,
+      "grad_norm": 0.2799666225910187,
+      "learning_rate": 3.2936507936507936e-05,
+      "loss": 0.0428,
+      "step": 3250
+    },
+    {
+      "epoch": 1.231203007518797,
+      "grad_norm": 0.30404964089393616,
+      "learning_rate": 3.2762461709830134e-05,
+      "loss": 0.0449,
+      "step": 3275
+    },
+    {
+      "epoch": 1.2406015037593985,
+      "grad_norm": 0.32838234305381775,
+      "learning_rate": 3.258841548315233e-05,
+      "loss": 0.0441,
+      "step": 3300
+    },
+    {
+      "epoch": 1.25,
+      "grad_norm": 0.41564425826072693,
+      "learning_rate": 3.241436925647452e-05,
+      "loss": 0.0406,
+      "step": 3325
+    },
+    {
+      "epoch": 1.2593984962406015,
+      "grad_norm": 0.6298181414604187,
+      "learning_rate": 3.2240323029796715e-05,
+      "loss": 0.0422,
+      "step": 3350
+    },
+    {
+      "epoch": 1.268796992481203,
+      "grad_norm": 0.4034960865974426,
+      "learning_rate": 3.206627680311891e-05,
+      "loss": 0.0399,
+      "step": 3375
+    },
+    {
+      "epoch": 1.2781954887218046,
+      "grad_norm": 0.24080845713615417,
+      "learning_rate": 3.1892230576441104e-05,
+      "loss": 0.042,
+      "step": 3400
+    },
+    {
+      "epoch": 1.287593984962406,
+      "grad_norm": 0.19642522931098938,
+      "learning_rate": 3.1718184349763296e-05,
+      "loss": 0.0454,
+      "step": 3425
+    },
+    {
+      "epoch": 1.2969924812030076,
+      "grad_norm": 0.45648419857025146,
+      "learning_rate": 3.1544138123085494e-05,
+      "loss": 0.0448,
+      "step": 3450
+    },
+    {
+      "epoch": 1.306390977443609,
+      "grad_norm": 0.28789225220680237,
+      "learning_rate": 3.1370091896407685e-05,
+      "loss": 0.0422,
+      "step": 3475
+    },
+    {
+      "epoch": 1.3157894736842106,
+      "grad_norm": 0.29442328214645386,
+      "learning_rate": 3.1196045669729884e-05,
+      "loss": 0.0419,
+      "step": 3500
+    },
+    {
+      "epoch": 1.3251879699248121,
+      "grad_norm": 0.23239637911319733,
+      "learning_rate": 3.1021999443052075e-05,
+      "loss": 0.0368,
+      "step": 3525
+    },
+    {
+      "epoch": 1.3345864661654137,
+      "grad_norm": 0.8643941283226013,
+      "learning_rate": 3.084795321637427e-05,
+      "loss": 0.0438,
+      "step": 3550
+    },
+    {
+      "epoch": 1.3439849624060152,
+      "grad_norm": 1.0415705442428589,
+      "learning_rate": 3.0673906989696465e-05,
+      "loss": 0.0371,
+      "step": 3575
+    },
+    {
+      "epoch": 1.3533834586466165,
+      "grad_norm": 0.36078059673309326,
+      "learning_rate": 3.0499860763018663e-05,
+      "loss": 0.0433,
+      "step": 3600
+    },
+    {
+      "epoch": 1.362781954887218,
+      "grad_norm": 0.564781129360199,
+      "learning_rate": 3.032581453634085e-05,
+      "loss": 0.0424,
+      "step": 3625
+    },
+    {
+      "epoch": 1.3721804511278195,
+      "grad_norm": 0.7637051343917847,
+      "learning_rate": 3.015176830966305e-05,
+      "loss": 0.0424,
+      "step": 3650
+    },
+    {
+      "epoch": 1.381578947368421,
+      "grad_norm": 0.5128487348556519,
+      "learning_rate": 2.9977722082985244e-05,
+      "loss": 0.0407,
+      "step": 3675
+    },
+    {
+      "epoch": 1.3909774436090225,
+      "grad_norm": 0.35614606738090515,
+      "learning_rate": 2.980367585630744e-05,
+      "loss": 0.0431,
+      "step": 3700
+    },
+    {
+      "epoch": 1.400375939849624,
+      "grad_norm": 0.25352317094802856,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.0418,
+      "step": 3725
+    },
+    {
+      "epoch": 1.4097744360902256,
+      "grad_norm": 0.42306146025657654,
+      "learning_rate": 2.9455583402951825e-05,
+      "loss": 0.0414,
+      "step": 3750
+    },
+    {
+      "epoch": 1.419172932330827,
+      "grad_norm": 0.8286581039428711,
+      "learning_rate": 2.928153717627402e-05,
+      "loss": 0.0432,
+      "step": 3775
+    },
+    {
+      "epoch": 1.4285714285714286,
+      "grad_norm": 0.6156339049339294,
+      "learning_rate": 2.9107490949596218e-05,
+      "loss": 0.0406,
+      "step": 3800
+    },
+    {
+      "epoch": 1.4379699248120301,
+      "grad_norm": 0.3396787643432617,
+      "learning_rate": 2.8933444722918406e-05,
+      "loss": 0.0446,
+      "step": 3825
+    },
+    {
+      "epoch": 1.4473684210526316,
+      "grad_norm": 0.32385748624801636,
+      "learning_rate": 2.8759398496240604e-05,
+      "loss": 0.0488,
+      "step": 3850
+    },
+    {
+      "epoch": 1.4567669172932332,
+      "grad_norm": 0.384922593832016,
+      "learning_rate": 2.85853522695628e-05,
+      "loss": 0.0413,
+      "step": 3875
+    },
+    {
+      "epoch": 1.4661654135338344,
+      "grad_norm": 0.8346031308174133,
+      "learning_rate": 2.8411306042884993e-05,
+      "loss": 0.0423,
+      "step": 3900
+    },
+    {
+      "epoch": 1.475563909774436,
+      "grad_norm": 0.6732437610626221,
+      "learning_rate": 2.8237259816207185e-05,
+      "loss": 0.0453,
+      "step": 3925
+    },
+    {
+      "epoch": 1.4849624060150375,
+      "grad_norm": 0.6928327679634094,
+      "learning_rate": 2.806321358952938e-05,
+      "loss": 0.0458,
+      "step": 3950
+    },
+    {
+      "epoch": 1.494360902255639,
+      "grad_norm": 0.3837633728981018,
+      "learning_rate": 2.7889167362851574e-05,
+      "loss": 0.0402,
+      "step": 3975
+    },
+    {
+      "epoch": 1.5037593984962405,
+      "grad_norm": 0.35443350672721863,
+      "learning_rate": 2.7715121136173773e-05,
+      "loss": 0.0416,
+      "step": 4000
+    },
+    {
+      "epoch": 1.513157894736842,
+      "grad_norm": 0.3493078649044037,
+      "learning_rate": 2.754803675856308e-05,
+      "loss": 0.0457,
+      "step": 4025
+    },
+    {
+      "epoch": 1.5225563909774436,
+      "grad_norm": 0.38165393471717834,
+      "learning_rate": 2.7373990531885268e-05,
+      "loss": 0.0458,
+      "step": 4050
+    },
+    {
+      "epoch": 1.531954887218045,
+      "grad_norm": 0.28904205560684204,
+      "learning_rate": 2.7199944305207466e-05,
+      "loss": 0.0456,
+      "step": 4075
+    },
+    {
+      "epoch": 1.5413533834586466,
+      "grad_norm": 0.27960851788520813,
+      "learning_rate": 2.702589807852966e-05,
+      "loss": 0.046,
+      "step": 4100
+    },
+    {
+      "epoch": 1.550751879699248,
+      "grad_norm": 0.5297873020172119,
+      "learning_rate": 2.6851851851851855e-05,
+      "loss": 0.045,
+      "step": 4125
+    },
+    {
+      "epoch": 1.5601503759398496,
+      "grad_norm": 0.5538225173950195,
+      "learning_rate": 2.6677805625174047e-05,
+      "loss": 0.0405,
+      "step": 4150
+    },
+    {
+      "epoch": 1.5695488721804511,
+      "grad_norm": 0.4728054702281952,
+      "learning_rate": 2.650375939849624e-05,
+      "loss": 0.0429,
+      "step": 4175
+    },
+    {
+      "epoch": 1.5789473684210527,
+      "grad_norm": 1.0343999862670898,
+      "learning_rate": 2.6329713171818436e-05,
+      "loss": 0.0474,
+      "step": 4200
+    },
+    {
+      "epoch": 1.5883458646616542,
+      "grad_norm": 0.5233461260795593,
+      "learning_rate": 2.6155666945140635e-05,
+      "loss": 0.0461,
+      "step": 4225
+    },
+    {
+      "epoch": 1.5977443609022557,
+      "grad_norm": 0.28031855821609497,
+      "learning_rate": 2.5981620718462823e-05,
+      "loss": 0.0428,
+      "step": 4250
+    },
+    {
+      "epoch": 1.6071428571428572,
+      "grad_norm": 0.30386048555374146,
+      "learning_rate": 2.580757449178502e-05,
+      "loss": 0.0404,
+      "step": 4275
+    },
+    {
+      "epoch": 1.6165413533834587,
+      "grad_norm": 0.2576204836368561,
+      "learning_rate": 2.5633528265107216e-05,
+      "loss": 0.042,
+      "step": 4300
+    },
+    {
+      "epoch": 1.6259398496240602,
+      "grad_norm": 0.2681669294834137,
+      "learning_rate": 2.545948203842941e-05,
+      "loss": 0.0429,
+      "step": 4325
+    },
+    {
+      "epoch": 1.6353383458646618,
+      "grad_norm": 0.19792746007442474,
+      "learning_rate": 2.5285435811751602e-05,
+      "loss": 0.0415,
+      "step": 4350
+    },
+    {
+      "epoch": 1.6447368421052633,
+      "grad_norm": 0.5182521939277649,
+      "learning_rate": 2.5111389585073797e-05,
+      "loss": 0.0476,
+      "step": 4375
+    },
+    {
+      "epoch": 1.6541353383458648,
+      "grad_norm": 0.587866485118866,
+      "learning_rate": 2.493734335839599e-05,
+      "loss": 0.044,
+      "step": 4400
+    },
+    {
+      "epoch": 1.6635338345864663,
+      "grad_norm": 0.3563426434993744,
+      "learning_rate": 2.4763297131718186e-05,
+      "loss": 0.0418,
+      "step": 4425
+    },
+    {
+      "epoch": 1.6729323308270678,
+      "grad_norm": 0.2907748818397522,
+      "learning_rate": 2.4589250905040378e-05,
+      "loss": 0.0438,
+      "step": 4450
+    },
+    {
+      "epoch": 1.6823308270676691,
+      "grad_norm": 0.58048015832901,
+      "learning_rate": 2.4415204678362576e-05,
+      "loss": 0.038,
+      "step": 4475
+    },
+    {
+      "epoch": 1.6917293233082706,
+      "grad_norm": 0.8364240527153015,
+      "learning_rate": 2.4241158451684767e-05,
+      "loss": 0.0433,
+      "step": 4500
+    },
+    {
+      "epoch": 1.7011278195488722,
+      "grad_norm": 1.025005578994751,
+      "learning_rate": 2.4067112225006962e-05,
+      "loss": 0.0391,
+      "step": 4525
+    },
+    {
+      "epoch": 1.7105263157894737,
+      "grad_norm": 0.513279139995575,
+      "learning_rate": 2.3893065998329157e-05,
+      "loss": 0.0461,
+      "step": 4550
+    },
+    {
+      "epoch": 1.7199248120300752,
+      "grad_norm": 0.7825498580932617,
+      "learning_rate": 2.371901977165135e-05,
+      "loss": 0.0397,
+      "step": 4575
+    },
+    {
+      "epoch": 1.7293233082706767,
+      "grad_norm": 0.3107648193836212,
+      "learning_rate": 2.3544973544973546e-05,
+      "loss": 0.0407,
+      "step": 4600
+    },
+    {
+      "epoch": 1.7387218045112782,
+      "grad_norm": 0.3225563168525696,
+      "learning_rate": 2.337092731829574e-05,
+      "loss": 0.0444,
+      "step": 4625
+    },
+    {
+      "epoch": 1.7481203007518797,
+      "grad_norm": 0.43894103169441223,
+      "learning_rate": 2.3196881091617932e-05,
+      "loss": 0.0431,
+      "step": 4650
+    },
+    {
+      "epoch": 1.7575187969924813,
+      "grad_norm": 0.33160117268562317,
+      "learning_rate": 2.302283486494013e-05,
+      "loss": 0.042,
+      "step": 4675
+    },
+    {
+      "epoch": 1.7669172932330826,
+      "grad_norm": 0.255582332611084,
+      "learning_rate": 2.2848788638262322e-05,
+      "loss": 0.0398,
+      "step": 4700
+    },
+    {
+      "epoch": 1.776315789473684,
+      "grad_norm": 0.2923487424850464,
+      "learning_rate": 2.2674742411584517e-05,
+      "loss": 0.0422,
+      "step": 4725
+    },
+    {
+      "epoch": 1.7857142857142856,
+      "grad_norm": 0.7796223759651184,
+      "learning_rate": 2.250069618490671e-05,
+      "loss": 0.0375,
+      "step": 4750
+    },
+    {
+      "epoch": 1.795112781954887,
+      "grad_norm": 0.4056415259838104,
+      "learning_rate": 2.2326649958228906e-05,
+      "loss": 0.0437,
+      "step": 4775
+    },
+    {
+      "epoch": 1.8045112781954886,
+      "grad_norm": 0.4501309394836426,
+      "learning_rate": 2.21526037315511e-05,
+      "loss": 0.0455,
+      "step": 4800
+    },
+    {
+      "epoch": 1.8139097744360901,
+      "grad_norm": 0.7883285880088806,
+      "learning_rate": 2.1978557504873296e-05,
+      "loss": 0.0443,
+      "step": 4825
+    },
+    {
+      "epoch": 1.8233082706766917,
+      "grad_norm": 0.33851534128189087,
+      "learning_rate": 2.1804511278195487e-05,
+      "loss": 0.0366,
+      "step": 4850
+    },
+    {
+      "epoch": 1.8327067669172932,
+      "grad_norm": 0.3407614529132843,
+      "learning_rate": 2.1630465051517686e-05,
+      "loss": 0.0415,
+      "step": 4875
+    },
+    {
+      "epoch": 1.8421052631578947,
+      "grad_norm": 0.4478450119495392,
+      "learning_rate": 2.1456418824839877e-05,
+      "loss": 0.0398,
+      "step": 4900
+    },
+    {
+      "epoch": 1.8515037593984962,
+      "grad_norm": 0.4571634531021118,
+      "learning_rate": 2.1282372598162072e-05,
+      "loss": 0.0382,
+      "step": 4925
+    },
+    {
+      "epoch": 1.8609022556390977,
+      "grad_norm": 1.4603149890899658,
+      "learning_rate": 2.1108326371484267e-05,
+      "loss": 0.0444,
+      "step": 4950
+    },
+    {
+      "epoch": 1.8703007518796992,
+      "grad_norm": 0.2934187054634094,
+      "learning_rate": 2.093428014480646e-05,
+      "loss": 0.0443,
+      "step": 4975
+    },
+    {
+      "epoch": 1.8796992481203008,
+      "grad_norm": 0.5655888319015503,
+      "learning_rate": 2.0760233918128656e-05,
+      "loss": 0.0431,
+      "step": 5000
+    },
+    {
+      "epoch": 1.8890977443609023,
+      "grad_norm": 0.9796953201293945,
+      "learning_rate": 2.058618769145085e-05,
+      "loss": 0.0424,
+      "step": 5025
+    },
+    {
+      "epoch": 1.8984962406015038,
+      "grad_norm": 0.6249701976776123,
+      "learning_rate": 2.0412141464773042e-05,
+      "loss": 0.0435,
+      "step": 5050
+    },
+    {
+      "epoch": 1.9078947368421053,
+      "grad_norm": 0.7648960947990417,
+      "learning_rate": 2.023809523809524e-05,
+      "loss": 0.0424,
+      "step": 5075
+    },
+    {
+      "epoch": 1.9172932330827068,
+      "grad_norm": 1.1739506721496582,
+      "learning_rate": 2.0064049011417432e-05,
+      "loss": 0.04,
+      "step": 5100
+    },
+    {
+      "epoch": 1.9266917293233083,
+      "grad_norm": 0.40099895000457764,
+      "learning_rate": 1.9890002784739627e-05,
+      "loss": 0.0426,
+      "step": 5125
+    },
+    {
+      "epoch": 1.9360902255639099,
+      "grad_norm": 0.443786084651947,
+      "learning_rate": 1.971595655806182e-05,
+      "loss": 0.0425,
+      "step": 5150
+    },
+    {
+      "epoch": 1.9454887218045114,
+      "grad_norm": 1.1793848276138306,
+      "learning_rate": 1.9541910331384016e-05,
+      "loss": 0.0412,
+      "step": 5175
+    },
+    {
+      "epoch": 1.954887218045113,
+      "grad_norm": 0.2933347225189209,
+      "learning_rate": 1.936786410470621e-05,
+      "loss": 0.0465,
+      "step": 5200
+    },
+    {
+      "epoch": 1.9642857142857144,
+      "grad_norm": 0.8869921565055847,
+      "learning_rate": 1.9193817878028406e-05,
+      "loss": 0.042,
+      "step": 5225
+    },
+    {
+      "epoch": 1.973684210526316,
+      "grad_norm": 0.6454533338546753,
+      "learning_rate": 1.9019771651350597e-05,
+      "loss": 0.0444,
+      "step": 5250
+    },
+    {
+      "epoch": 1.9830827067669174,
+      "grad_norm": 0.4456064999103546,
+      "learning_rate": 1.8845725424672795e-05,
+      "loss": 0.0398,
+      "step": 5275
+    },
+    {
+      "epoch": 1.9924812030075187,
+      "grad_norm": 0.7413787245750427,
+      "learning_rate": 1.8671679197994987e-05,
+      "loss": 0.0458,
+      "step": 5300
+    },
+    {
+      "epoch": 2.0,
+      "eval_explained_variance": 0.23339885473251343,
+      "eval_loss": 0.0432429239153862,
+      "eval_mae": 0.160363107919693,
+      "eval_mse": 0.043248895555734634,
+      "eval_r2": 0.21744608879089355,
+      "eval_rmse": 0.20796368807014035,
+      "eval_runtime": 5.0094,
+      "eval_samples_per_second": 4247.455,
+      "eval_steps_per_second": 66.476,
+      "step": 5320
+    },
+    {
+      "epoch": 2.0018796992481205,
+      "grad_norm": 0.30451804399490356,
+      "learning_rate": 1.849763297131718e-05,
+      "loss": 0.0383,
+      "step": 5325
+    },
+    {
+      "epoch": 2.011278195488722,
+      "grad_norm": 0.3505537211894989,
+      "learning_rate": 1.8323586744639376e-05,
+      "loss": 0.0372,
+      "step": 5350
+    },
+    {
+      "epoch": 2.0206766917293235,
+      "grad_norm": 0.3725011646747589,
+      "learning_rate": 1.814954051796157e-05,
+      "loss": 0.0395,
+      "step": 5375
+    },
+    {
+      "epoch": 2.030075187969925,
+      "grad_norm": 0.45913949608802795,
+      "learning_rate": 1.7975494291283766e-05,
+      "loss": 0.0378,
+      "step": 5400
+    },
+    {
+      "epoch": 2.039473684210526,
+      "grad_norm": 0.24697057902812958,
+      "learning_rate": 1.780144806460596e-05,
+      "loss": 0.0359,
+      "step": 5425
+    },
+    {
+      "epoch": 2.0488721804511276,
+      "grad_norm": 0.32664352655410767,
+      "learning_rate": 1.7627401837928152e-05,
+      "loss": 0.0342,
+      "step": 5450
+    },
+    {
+      "epoch": 2.058270676691729,
+      "grad_norm": 0.3500126600265503,
+      "learning_rate": 1.745335561125035e-05,
+      "loss": 0.0375,
+      "step": 5475
+    },
+    {
+      "epoch": 2.0676691729323307,
+      "grad_norm": 0.2665722668170929,
+      "learning_rate": 1.7279309384572542e-05,
+      "loss": 0.0384,
+      "step": 5500
+    },
+    {
+      "epoch": 2.077067669172932,
+      "grad_norm": 0.2083020657300949,
+      "learning_rate": 1.7105263157894737e-05,
+      "loss": 0.0377,
+      "step": 5525
+    },
+    {
+      "epoch": 2.0864661654135337,
+      "grad_norm": 0.5617522597312927,
+      "learning_rate": 1.693121693121693e-05,
+      "loss": 0.0389,
+      "step": 5550
+    },
+    {
+      "epoch": 2.095864661654135,
+      "grad_norm": 0.3643837869167328,
+      "learning_rate": 1.6757170704539126e-05,
+      "loss": 0.0349,
+      "step": 5575
+    },
+    {
+      "epoch": 2.1052631578947367,
+      "grad_norm": 0.8100130558013916,
+      "learning_rate": 1.658312447786132e-05,
+      "loss": 0.0356,
+      "step": 5600
+    },
+    {
+      "epoch": 2.1146616541353382,
+      "grad_norm": 0.24865201115608215,
+      "learning_rate": 1.6409078251183516e-05,
+      "loss": 0.0352,
+      "step": 5625
+    },
+    {
+      "epoch": 2.1240601503759398,
+      "grad_norm": 1.1824196577072144,
+      "learning_rate": 1.6235032024505707e-05,
+      "loss": 0.0366,
+      "step": 5650
+    },
+    {
+      "epoch": 2.1334586466165413,
+      "grad_norm": 0.31902220845222473,
+      "learning_rate": 1.6060985797827905e-05,
+      "loss": 0.0362,
+      "step": 5675
+    },
+    {
+      "epoch": 2.142857142857143,
+      "grad_norm": 0.8043758869171143,
+      "learning_rate": 1.5886939571150097e-05,
+      "loss": 0.0376,
+      "step": 5700
+    },
+    {
+      "epoch": 2.1522556390977443,
+      "grad_norm": 0.40435153245925903,
+      "learning_rate": 1.571289334447229e-05,
+      "loss": 0.0384,
+      "step": 5725
+    },
+    {
+      "epoch": 2.161654135338346,
+      "grad_norm": 0.4308648407459259,
+      "learning_rate": 1.5538847117794486e-05,
+      "loss": 0.0369,
+      "step": 5750
+    },
+    {
+      "epoch": 2.1710526315789473,
+      "grad_norm": 0.4685553312301636,
+      "learning_rate": 1.536480089111668e-05,
+      "loss": 0.0376,
+      "step": 5775
+    },
+    {
+      "epoch": 2.180451127819549,
+      "grad_norm": 0.9758443832397461,
+      "learning_rate": 1.5190754664438874e-05,
+      "loss": 0.0381,
+      "step": 5800
+    },
+    {
+      "epoch": 2.1898496240601504,
+      "grad_norm": 0.44810280203819275,
+      "learning_rate": 1.501670843776107e-05,
+      "loss": 0.0408,
+      "step": 5825
+    },
+    {
+      "epoch": 2.199248120300752,
+      "grad_norm": 0.3626968264579773,
+      "learning_rate": 1.4842662211083264e-05,
+      "loss": 0.0345,
+      "step": 5850
+    },
+    {
+      "epoch": 2.2086466165413534,
+      "grad_norm": 0.4912227690219879,
+      "learning_rate": 1.4668615984405459e-05,
+      "loss": 0.0388,
+      "step": 5875
+    },
+    {
+      "epoch": 2.218045112781955,
+      "grad_norm": 0.39659154415130615,
+      "learning_rate": 1.4494569757727652e-05,
+      "loss": 0.0346,
+      "step": 5900
+    },
+    {
+      "epoch": 2.2274436090225564,
+      "grad_norm": 0.7727639079093933,
+      "learning_rate": 1.4320523531049848e-05,
+      "loss": 0.0387,
+      "step": 5925
+    },
+    {
+      "epoch": 2.236842105263158,
+      "grad_norm": 0.4822790026664734,
+      "learning_rate": 1.4146477304372041e-05,
+      "loss": 0.0367,
+      "step": 5950
+    },
+    {
+      "epoch": 2.2462406015037595,
+      "grad_norm": 0.955757737159729,
+      "learning_rate": 1.3972431077694236e-05,
+      "loss": 0.0359,
+      "step": 5975
+    },
+    {
+      "epoch": 2.255639097744361,
+      "grad_norm": 0.39219462871551514,
+      "learning_rate": 1.3798384851016429e-05,
+      "loss": 0.04,
+      "step": 6000
+    },
+    {
+      "epoch": 2.2650375939849625,
+      "grad_norm": 0.3432631194591522,
+      "learning_rate": 1.3624338624338626e-05,
+      "loss": 0.0384,
+      "step": 6025
+    },
+    {
+      "epoch": 2.274436090225564,
+      "grad_norm": 0.47153186798095703,
+      "learning_rate": 1.3450292397660819e-05,
+      "loss": 0.0385,
+      "step": 6050
+    },
+    {
+      "epoch": 2.2838345864661656,
+      "grad_norm": 0.3772079348564148,
+      "learning_rate": 1.3276246170983015e-05,
+      "loss": 0.0356,
+      "step": 6075
+    },
+    {
+      "epoch": 2.293233082706767,
+      "grad_norm": 0.4212327301502228,
+      "learning_rate": 1.3102199944305207e-05,
+      "loss": 0.0386,
+      "step": 6100
+    },
+    {
+      "epoch": 2.3026315789473686,
+      "grad_norm": 0.4279068410396576,
+      "learning_rate": 1.2928153717627403e-05,
+      "loss": 0.0341,
+      "step": 6125
+    },
+    {
+      "epoch": 2.31203007518797,
+      "grad_norm": 0.3151835501194,
+      "learning_rate": 1.276106934001671e-05,
+      "loss": 0.0376,
+      "step": 6150
+    },
+    {
+      "epoch": 2.3214285714285716,
+      "grad_norm": 0.7829068303108215,
+      "learning_rate": 1.2587023113338903e-05,
+      "loss": 0.0373,
+      "step": 6175
+    },
+    {
+      "epoch": 2.3308270676691727,
+      "grad_norm": 0.2983991801738739,
+      "learning_rate": 1.2412976886661096e-05,
+      "loss": 0.0376,
+      "step": 6200
+    },
+    {
+      "epoch": 2.340225563909774,
+      "grad_norm": 0.36468955874443054,
+      "learning_rate": 1.2238930659983291e-05,
+      "loss": 0.0368,
+      "step": 6225
+    },
+    {
+      "epoch": 2.3496240601503757,
+      "grad_norm": 0.2933436632156372,
+      "learning_rate": 1.2064884433305486e-05,
+      "loss": 0.0373,
+      "step": 6250
+    },
+    {
+      "epoch": 2.3590225563909772,
+      "grad_norm": 0.9246677160263062,
+      "learning_rate": 1.189083820662768e-05,
+      "loss": 0.038,
+      "step": 6275
+    },
+    {
+      "epoch": 2.3684210526315788,
+      "grad_norm": 0.7416250109672546,
+      "learning_rate": 1.1716791979949874e-05,
+      "loss": 0.0379,
+      "step": 6300
+    },
+    {
+      "epoch": 2.3778195488721803,
+      "grad_norm": 0.28008949756622314,
+      "learning_rate": 1.1542745753272069e-05,
+      "loss": 0.0342,
+      "step": 6325
+    },
+    {
+      "epoch": 2.387218045112782,
+      "grad_norm": 0.36140310764312744,
+      "learning_rate": 1.1368699526594263e-05,
+      "loss": 0.0358,
+      "step": 6350
+    },
+    {
+      "epoch": 2.3966165413533833,
+      "grad_norm": 0.2620641887187958,
+      "learning_rate": 1.1194653299916458e-05,
+      "loss": 0.0388,
+      "step": 6375
+    },
+    {
+      "epoch": 2.406015037593985,
+      "grad_norm": 0.48447370529174805,
+      "learning_rate": 1.1020607073238653e-05,
+      "loss": 0.0377,
+      "step": 6400
+    },
+    {
+      "epoch": 2.4154135338345863,
+      "grad_norm": 0.718144416809082,
+      "learning_rate": 1.0846560846560846e-05,
+      "loss": 0.0394,
+      "step": 6425
+    },
+    {
+      "epoch": 2.424812030075188,
+      "grad_norm": 1.0914640426635742,
+      "learning_rate": 1.067251461988304e-05,
+      "loss": 0.0352,
+      "step": 6450
+    },
+    {
+      "epoch": 2.4342105263157894,
+      "grad_norm": 0.997230589389801,
+      "learning_rate": 1.0498468393205236e-05,
+      "loss": 0.0395,
+      "step": 6475
+    },
+    {
+      "epoch": 2.443609022556391,
+      "grad_norm": 0.3665817975997925,
+      "learning_rate": 1.032442216652743e-05,
+      "loss": 0.0378,
+      "step": 6500
+    },
+    {
+      "epoch": 2.4530075187969924,
+      "grad_norm": 0.35913845896720886,
+      "learning_rate": 1.0150375939849624e-05,
+      "loss": 0.0382,
+      "step": 6525
+    },
+    {
+      "epoch": 2.462406015037594,
+      "grad_norm": 0.9548642635345459,
+      "learning_rate": 9.976329713171818e-06,
+      "loss": 0.0386,
+      "step": 6550
+    },
+    {
+      "epoch": 2.4718045112781954,
+      "grad_norm": 1.2121777534484863,
+      "learning_rate": 9.802283486494013e-06,
+      "loss": 0.0392,
+      "step": 6575
+    },
+    {
+      "epoch": 2.481203007518797,
+      "grad_norm": 0.568284809589386,
+      "learning_rate": 9.628237259816208e-06,
+      "loss": 0.0344,
+      "step": 6600
+    },
+    {
+      "epoch": 2.4906015037593985,
+      "grad_norm": 0.40139758586883545,
+      "learning_rate": 9.454191033138401e-06,
+      "loss": 0.0377,
+      "step": 6625
+    },
+    {
+      "epoch": 2.5,
+      "grad_norm": 0.5859258770942688,
+      "learning_rate": 9.280144806460596e-06,
+      "loss": 0.0342,
+      "step": 6650
+    },
+    {
+      "epoch": 2.5093984962406015,
+      "grad_norm": 0.4431360363960266,
+      "learning_rate": 9.10609857978279e-06,
+      "loss": 0.0398,
+      "step": 6675
+    },
+    {
+      "epoch": 2.518796992481203,
+      "grad_norm": 0.2714449465274811,
+      "learning_rate": 8.932052353104985e-06,
+      "loss": 0.0394,
+      "step": 6700
+    },
+    {
+      "epoch": 2.5281954887218046,
+      "grad_norm": 0.29640740156173706,
+      "learning_rate": 8.758006126427178e-06,
+      "loss": 0.0365,
+      "step": 6725
+    },
+    {
+      "epoch": 2.537593984962406,
+      "grad_norm": 0.7424734830856323,
+      "learning_rate": 8.583959899749373e-06,
+      "loss": 0.037,
+      "step": 6750
+    },
+    {
+      "epoch": 2.5469924812030076,
+      "grad_norm": 0.49265924096107483,
+      "learning_rate": 8.409913673071568e-06,
+      "loss": 0.0357,
+      "step": 6775
+    },
+    {
+      "epoch": 2.556390977443609,
+      "grad_norm": 0.8623179197311401,
+      "learning_rate": 8.235867446393763e-06,
+      "loss": 0.0388,
+      "step": 6800
+    },
+    {
+      "epoch": 2.5657894736842106,
+      "grad_norm": 0.4578423500061035,
+      "learning_rate": 8.061821219715956e-06,
+      "loss": 0.0395,
+      "step": 6825
+    },
+    {
+      "epoch": 2.575187969924812,
+      "grad_norm": 0.6307206749916077,
+      "learning_rate": 7.88777499303815e-06,
+      "loss": 0.033,
+      "step": 6850
+    },
+    {
+      "epoch": 2.5845864661654137,
+      "grad_norm": 0.49787333607673645,
+      "learning_rate": 7.713728766360345e-06,
+      "loss": 0.0398,
+      "step": 6875
+    },
+    {
+      "epoch": 2.593984962406015,
+      "grad_norm": 0.5026729106903076,
+      "learning_rate": 7.5396825396825394e-06,
+      "loss": 0.0374,
+      "step": 6900
+    },
+    {
+      "epoch": 2.6033834586466167,
+      "grad_norm": 0.24913954734802246,
+      "learning_rate": 7.365636313004734e-06,
+      "loss": 0.0372,
+      "step": 6925
+    },
+    {
+      "epoch": 2.612781954887218,
+      "grad_norm": 1.084121823310852,
+      "learning_rate": 7.191590086326928e-06,
+      "loss": 0.0407,
+      "step": 6950
+    },
+    {
+      "epoch": 2.6221804511278197,
+      "grad_norm": 0.9638610482215881,
+      "learning_rate": 7.017543859649123e-06,
+      "loss": 0.04,
+      "step": 6975
+    },
+    {
+      "epoch": 2.6315789473684212,
+      "grad_norm": 0.5844775438308716,
+      "learning_rate": 6.843497632971317e-06,
+      "loss": 0.0366,
+      "step": 7000
+    },
+    {
+      "epoch": 2.6409774436090228,
+      "grad_norm": 0.5980415940284729,
+      "learning_rate": 6.669451406293512e-06,
+      "loss": 0.0378,
+      "step": 7025
+    },
+    {
+      "epoch": 2.6503759398496243,
+      "grad_norm": 0.40126296877861023,
+      "learning_rate": 6.495405179615706e-06,
+      "loss": 0.0355,
+      "step": 7050
+    },
+    {
+      "epoch": 2.659774436090226,
+      "grad_norm": 0.5292906761169434,
+      "learning_rate": 6.3213589529379004e-06,
+      "loss": 0.0342,
+      "step": 7075
+    },
+    {
+      "epoch": 2.6691729323308273,
+      "grad_norm": 0.5242503881454468,
+      "learning_rate": 6.147312726260094e-06,
+      "loss": 0.041,
+      "step": 7100
+    },
+    {
+      "epoch": 2.678571428571429,
+      "grad_norm": 0.6571367383003235,
+      "learning_rate": 5.973266499582289e-06,
+      "loss": 0.0342,
+      "step": 7125
+    },
+    {
+      "epoch": 2.6879699248120303,
+      "grad_norm": 0.3011007606983185,
+      "learning_rate": 5.799220272904483e-06,
+      "loss": 0.0371,
+      "step": 7150
+    },
+    {
+      "epoch": 2.6973684210526314,
+      "grad_norm": 0.9834239482879639,
+      "learning_rate": 5.625174046226678e-06,
+      "loss": 0.0401,
+      "step": 7175
+    },
+    {
+      "epoch": 2.706766917293233,
+      "grad_norm": 0.49396082758903503,
+      "learning_rate": 5.451127819548872e-06,
+      "loss": 0.0377,
+      "step": 7200
+    },
+    {
+      "epoch": 2.7161654135338344,
+      "grad_norm": 0.7074937224388123,
+      "learning_rate": 5.277081592871067e-06,
+      "loss": 0.0342,
+      "step": 7225
+    },
+    {
+      "epoch": 2.725563909774436,
+      "grad_norm": 0.6037689447402954,
+      "learning_rate": 5.103035366193261e-06,
+      "loss": 0.0353,
+      "step": 7250
+    },
+    {
+      "epoch": 2.7349624060150375,
+      "grad_norm": 0.9242851734161377,
+      "learning_rate": 4.928989139515455e-06,
+      "loss": 0.0378,
+      "step": 7275
+    },
+    {
+      "epoch": 2.744360902255639,
+      "grad_norm": 0.3472757339477539,
+      "learning_rate": 4.754942912837649e-06,
+      "loss": 0.0335,
+      "step": 7300
+    },
+    {
+      "epoch": 2.7537593984962405,
+      "grad_norm": 0.5755854845046997,
+      "learning_rate": 4.580896686159844e-06,
+      "loss": 0.032,
+      "step": 7325
+    },
+    {
+      "epoch": 2.763157894736842,
+      "grad_norm": 0.37826091051101685,
+      "learning_rate": 4.406850459482038e-06,
+      "loss": 0.0354,
+      "step": 7350
+    },
+    {
+      "epoch": 2.7725563909774436,
+      "grad_norm": 0.5390151143074036,
+      "learning_rate": 4.232804232804233e-06,
+      "loss": 0.0341,
+      "step": 7375
+    },
+    {
+      "epoch": 2.781954887218045,
+      "grad_norm": 0.28750574588775635,
+      "learning_rate": 4.058758006126427e-06,
+      "loss": 0.0372,
+      "step": 7400
+    },
+    {
+      "epoch": 2.7913533834586466,
+      "grad_norm": 0.6178227663040161,
+      "learning_rate": 3.884711779448622e-06,
+      "loss": 0.0353,
+      "step": 7425
+    },
+    {
+      "epoch": 2.800751879699248,
+      "grad_norm": 0.3084428608417511,
+      "learning_rate": 3.710665552770816e-06,
+      "loss": 0.0371,
+      "step": 7450
+    },
+    {
+      "epoch": 2.8101503759398496,
+      "grad_norm": 0.3735020160675049,
+      "learning_rate": 3.5366193260930103e-06,
+      "loss": 0.0365,
+      "step": 7475
+    },
+    {
+      "epoch": 2.819548872180451,
+      "grad_norm": 0.3113434910774231,
+      "learning_rate": 3.3625730994152047e-06,
+      "loss": 0.0395,
+      "step": 7500
+    },
+    {
+      "epoch": 2.8289473684210527,
+      "grad_norm": 0.7996252775192261,
+      "learning_rate": 3.188526872737399e-06,
+      "loss": 0.0382,
+      "step": 7525
+    },
+    {
+      "epoch": 2.838345864661654,
+      "grad_norm": 0.3369075357913971,
+      "learning_rate": 3.0144806460595934e-06,
+      "loss": 0.0338,
+      "step": 7550
+    },
+    {
+      "epoch": 2.8477443609022557,
+      "grad_norm": 0.326662540435791,
+      "learning_rate": 2.8404344193817878e-06,
+      "loss": 0.0333,
+      "step": 7575
+    },
+    {
+      "epoch": 2.857142857142857,
+      "grad_norm": 0.47569015622138977,
+      "learning_rate": 2.666388192703982e-06,
+      "loss": 0.0361,
+      "step": 7600
+    },
+    {
+      "epoch": 2.8665413533834587,
+      "grad_norm": 0.489814430475235,
+      "learning_rate": 2.4923419660261765e-06,
+      "loss": 0.0354,
+      "step": 7625
+    },
+    {
+      "epoch": 2.8759398496240602,
+      "grad_norm": 0.3687739372253418,
+      "learning_rate": 2.318295739348371e-06,
+      "loss": 0.0386,
+      "step": 7650
+    },
+    {
+      "epoch": 2.8853383458646618,
+      "grad_norm": 0.8088187575340271,
+      "learning_rate": 2.1442495126705653e-06,
+      "loss": 0.0367,
+      "step": 7675
+    },
+    {
+      "epoch": 2.8947368421052633,
+      "grad_norm": 0.41726213693618774,
+      "learning_rate": 1.9702032859927596e-06,
+      "loss": 0.0364,
+      "step": 7700
+    },
+    {
+      "epoch": 2.904135338345865,
+      "grad_norm": 0.3655787408351898,
+      "learning_rate": 1.796157059314954e-06,
+      "loss": 0.0363,
+      "step": 7725
+    },
+    {
+      "epoch": 2.9135338345864663,
+      "grad_norm": 0.49075835943222046,
+      "learning_rate": 1.6221108326371484e-06,
+      "loss": 0.0349,
+      "step": 7750
+    },
+    {
+      "epoch": 2.922932330827068,
+      "grad_norm": 0.3516837954521179,
+      "learning_rate": 1.4480646059593427e-06,
+      "loss": 0.0354,
+      "step": 7775
+    },
+    {
+      "epoch": 2.932330827067669,
+      "grad_norm": 0.40250521898269653,
+      "learning_rate": 1.274018379281537e-06,
+      "loss": 0.0412,
+      "step": 7800
+    },
+    {
+      "epoch": 2.9417293233082704,
+      "grad_norm": 0.33906999230384827,
+      "learning_rate": 1.0999721526037315e-06,
+      "loss": 0.0375,
+      "step": 7825
+    },
+    {
+      "epoch": 2.951127819548872,
+      "grad_norm": 0.37088847160339355,
+      "learning_rate": 9.259259259259259e-07,
+      "loss": 0.039,
+      "step": 7850
+    },
+    {
+      "epoch": 2.9605263157894735,
+      "grad_norm": 0.2640984058380127,
+      "learning_rate": 7.518796992481203e-07,
+      "loss": 0.0359,
+      "step": 7875
+    },
+    {
+      "epoch": 2.969924812030075,
+      "grad_norm": 0.9323065280914307,
+      "learning_rate": 5.778334725703147e-07,
+      "loss": 0.0378,
+      "step": 7900
+    },
+    {
+      "epoch": 2.9793233082706765,
+      "grad_norm": 0.6576139330863953,
+      "learning_rate": 4.0378724589250904e-07,
+      "loss": 0.0362,
+      "step": 7925
+    },
+    {
+      "epoch": 2.988721804511278,
+      "grad_norm": 0.7086507081985474,
+      "learning_rate": 2.2974101921470343e-07,
+      "loss": 0.0306,
+      "step": 7950
+    },
+    {
+      "epoch": 2.9981203007518795,
+      "grad_norm": 0.9082393050193787,
+      "learning_rate": 5.56947925368978e-08,
+      "loss": 0.0372,
+      "step": 7975
+    },
+    {
+      "epoch": 3.0,
+      "eval_explained_variance": 0.23569148778915405,
+      "eval_loss": 0.04224373400211334,
+      "eval_mae": 0.15929067134857178,
+      "eval_mse": 0.04224866256117821,
+      "eval_r2": 0.23554444313049316,
+      "eval_rmse": 0.20554479453680702,
+      "eval_runtime": 5.0467,
+      "eval_samples_per_second": 4216.047,
+      "eval_steps_per_second": 65.984,
+      "step": 7980
+    }
+  ],
+  "logging_steps": 25,
+  "max_steps": 7980,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 5,
+        "early_stopping_threshold": 0.01
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 2
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 8456567720509440.0,
+  "train_batch_size": 8,
+  "trial_name": null,
+  "trial_params": null
+}

checkpoint-7980/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:414540c031bffc02f1019ad7982fdf75ad9763d9b08b73d3be8ef2c8b11e0e01
+size 5368

config.json

@@ -0,0 +1,32 @@
+{
+  "_name_or_path": "distilbert/distilbert-base-uncased",
+  "_num_labels": 1,
+  "activation": "gelu",
+  "architectures": [
+    "DistilBertForSequenceClassification"
+  ],
+  "attention_dropout": 0.1,
+  "dim": 768,
+  "dropout": 0.1,
+  "hidden_dim": 3072,
+  "id2label": {
+    "0": "target"
+  },
+  "initializer_range": 0.02,
+  "label2id": {
+    "target": 0
+  },
+  "max_position_embeddings": 512,
+  "model_type": "distilbert",
+  "n_heads": 12,
+  "n_layers": 6,
+  "pad_token_id": 0,
+  "problem_type": "regression",
+  "qa_dropout": 0.1,
+  "seq_classif_dropout": 0.2,
+  "sinusoidal_pos_embds": false,
+  "tie_weights_": true,
+  "torch_dtype": "float32",
+  "transformers_version": "4.48.0",
+  "vocab_size": 30522
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7eebdfe40823a90e15ce6d5701dbe3da53f026b8ce64e6c583ed64be7aa1150d
+size 267829484

runs/Feb22_17-34-28_r-samtuckervegan-autotrain-4xl4-vsnvk6q0-6181f-d0bud/events.out.tfevents.1740245669.r-samtuckervegan-autotrain-4xl4-vsnvk6q0-6181f-d0bud.201.0

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5667c72d991ec974a0c3fd8e441de75cc170f59d62e86c87616be644ce333db7
-size 58816
+oid sha256:eca971b07a061729a987a2eb764e26e9b1ed026f06227aa3a6587e1758c8ae03
+size 74250

runs/Feb22_17-34-28_r-samtuckervegan-autotrain-4xl4-vsnvk6q0-6181f-d0bud/events.out.tfevents.1740246447.r-samtuckervegan-autotrain-4xl4-vsnvk6q0-6181f-d0bud.201.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3ef271a3ffb344ad06bb55f5d4f56b001741df8dd744d62eceeac609a14ba5a6
+size 609

special_tokens_map.json

@@ -0,0 +1,7 @@
+{
+  "cls_token": "[CLS]",
+  "mask_token": "[MASK]",
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "unk_token": "[UNK]"
+}

tokenizer_config.json

@@ -0,0 +1,56 @@
+{
+  "added_tokens_decoder": {
+    "0": {
+      "content": "[PAD]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "100": {
+      "content": "[UNK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "101": {
+      "content": "[CLS]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "102": {
+      "content": "[SEP]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    },
+    "103": {
+      "content": "[MASK]",
+      "lstrip": false,
+      "normalized": false,
+      "rstrip": false,
+      "single_word": false,
+      "special": true
+    }
+  },
+  "clean_up_tokenization_spaces": false,
+  "cls_token": "[CLS]",
+  "do_lower_case": true,
+  "extra_special_tokens": {},
+  "mask_token": "[MASK]",
+  "model_max_length": 512,
+  "pad_token": "[PAD]",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "DistilBertTokenizer",
+  "unk_token": "[UNK]"
+}

training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:414540c031bffc02f1019ad7982fdf75ad9763d9b08b73d3be8ef2c8b11e0e01
+size 5368

training_params.json

@@ -0,0 +1,30 @@
+{
+    "data_path": "samtuckervegan/relevancy",
+    "model": "distilbert/distilbert-base-uncased",
+    "lr": 5e-05,
+    "epochs": 3,
+    "max_seq_length": 128,
+    "batch_size": 8,
+    "warmup_ratio": 0.1,
+    "gradient_accumulation": 1,
+    "optimizer": "adamw_torch",
+    "scheduler": "linear",
+    "weight_decay": 0.0,
+    "max_grad_norm": 1.0,
+    "seed": 42,
+    "train_split": "train",
+    "valid_split": "test",
+    "text_column": "text",
+    "target_column": "target",
+    "logging_steps": -1,
+    "project_name": "relevancy",
+    "auto_find_batch_size": false,
+    "mixed_precision": "fp16",
+    "save_total_limit": 1,
+    "push_to_hub": true,
+    "eval_strategy": "epoch",
+    "username": "samtuckervegan",
+    "log": "tensorboard",
+    "early_stopping_patience": 5,
+    "early_stopping_threshold": 0.01
+}