End of training

README.md

@@ -1,6 +1,6 @@
 ---
 base_model: LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct
-datasets: testing_layer_28_Matrix_mlp.c_proj
+datasets: testing_layer_28_Matrix_attn.attention.v_proj
 library_name: transformers
 model_name: output_exaone_tucker_rank16_newdata_b1_g2_gradient2_e1_b512
 tags:
@@ -12,7 +12,7 @@ licence: license
 
 # Model Card for output_exaone_tucker_rank16_newdata_b1_g2_gradient2_e1_b512
 
-This model is a fine-tuned version of [LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct](https://huggingface.co/LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct) on the [testing_layer_28_Matrix_mlp.c_proj](https://huggingface.co/datasets/testing_layer_28_Matrix_mlp.c_proj) dataset.
+This model is a fine-tuned version of [LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct](https://huggingface.co/LGAI-EXAONE/EXAONE-3.5-2.4B-Instruct) on the [testing_layer_28_Matrix_attn.attention.v_proj](https://huggingface.co/datasets/testing_layer_28_Matrix_attn.attention.v_proj) dataset.
 It has been trained using [TRL](https://github.com/huggingface/trl).
 
 ## Quick start

all_results.json

@@ -1,15 +1,15 @@
 {
-    "eval_accuracy": 0.8091990942751283,
-    "eval_loss": 1.6708984375,
-    "eval_runtime": 934.1658,
+    "eval_accuracy": 0.8090387906535197,
+    "eval_loss": 1.6748046875,
+    "eval_runtime": 942.4913,
     "eval_samples": 25971,
-    "eval_samples_per_second": 27.801,
-    "eval_steps_per_second": 13.901,
-    "perplexity": 5.31694259254287,
-    "total_flos": 9.078073041438638e+17,
-    "train_loss": 1.6789688296530434,
-    "train_runtime": 12675.9905,
+    "eval_samples_per_second": 27.556,
+    "eval_steps_per_second": 13.778,
+    "perplexity": 5.33775251746843,
+    "total_flos": 9.005567628877496e+17,
+    "train_loss": 1.67220352899737,
+    "train_runtime": 13481.0993,
     "train_samples": 294393,
-    "train_samples_per_second": 23.224,
-    "train_steps_per_second": 5.806
+    "train_samples_per_second": 21.837,
+    "train_steps_per_second": 5.459
 }

eval_layer_index_28_mat_attn.attention.v_proj_results.json

@@ -0,0 +1,9 @@
+{
+    "eval_accuracy": 0.8090387906535197,
+    "eval_loss": 1.6748046875,
+    "eval_runtime": 942.4913,
+    "eval_samples": 25971,
+    "eval_samples_per_second": 27.556,
+    "eval_steps_per_second": 13.778,
+    "perplexity": 5.33775251746843
+}

model-00001-of-00002.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:c4a4e85ebd672ede1b25882491b132e29791e040f5587628d0261e4debb03cf3
-size 4560124544
+oid sha256:382608745c31296f15dd7c33b8132dafc9b462e896e2368ae8b2bca668f47ec5
+size 4527892128

model.safetensors.index.json

@@ -1,6 +1,6 @@
 {
   "metadata": {
-    "total_size": 5084379136
+    "total_size": 5052145664
   },
   "weight_map": {
     "lm_head.weight": "model-00002-of-00002.safetensors",
@@ -193,10 +193,18 @@
     "transformer.h.27.mlp.c_fc_0.weight": "model-00001-of-00002.safetensors",
     "transformer.h.27.mlp.c_fc_1.weight": "model-00001-of-00002.safetensors",
     "transformer.h.27.mlp.c_proj.weight": "model-00001-of-00002.safetensors",
-    "transformer.h.28.attn.attention.k_proj.weight": "model-00001-of-00002.safetensors",
-    "transformer.h.28.attn.attention.out_proj.weight": "model-00001-of-00002.safetensors",
-    "transformer.h.28.attn.attention.q_proj.weight": "model-00001-of-00002.safetensors",
-    "transformer.h.28.attn.attention.v_proj.weight": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.k_proj.tucker_core": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.k_proj.tucker_factors.0": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.k_proj.tucker_factors.1": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.out_proj.tucker_core": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.out_proj.tucker_factors.0": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.out_proj.tucker_factors.1": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.q_proj.tucker_core": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.q_proj.tucker_factors.0": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.q_proj.tucker_factors.1": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.v_proj.tucker_core": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.v_proj.tucker_factors.0": "model-00001-of-00002.safetensors",
+    "transformer.h.28.attn.attention.v_proj.tucker_factors.1": "model-00001-of-00002.safetensors",
     "transformer.h.28.ln_1.weight": "model-00001-of-00002.safetensors",
     "transformer.h.28.ln_2.weight": "model-00001-of-00002.safetensors",
     "transformer.h.28.mlp.c_fc_0.tucker_core": "model-00001-of-00002.safetensors",

runs/Mar13_16-43-27_demo2/events.out.tfevents.1741954545.demo2.205134.6

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c1232ec953a5b0764ae4849dde1e7af63915bf537f9a59d8030ff08d639f6c3d
+size 47428

runs/Mar13_16-43-27_demo2/events.out.tfevents.1741968998.demo2.205134.7

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4676f1da5149dc74dab578345a27f0586f0af8cfbfbdadfdaaedd90cb5928832
+size 481

train_layer_index_28_mat_attn.attention.v_proj_results.json

@@ -0,0 +1,8 @@
+{
+    "total_flos": 9.005567628877496e+17,
+    "train_loss": 1.67220352899737,
+    "train_runtime": 13481.0993,
+    "train_samples": 294393,
+    "train_samples_per_second": 21.837,
+    "train_steps_per_second": 5.459
+}

trainer_state.json

@@ -10,1189 +10,1189 @@
   "log_history": [
     {
       "epoch": 0.006793616717732019,
-      "grad_norm": 0.4056946635246277,
-      "learning_rate": 0.000496623549552977,
-      "loss": 1.7688,
-      "mean_token_accuracy": 0.5910043245553971,
+      "grad_norm": 0.203684464097023,
+      "learning_rate": 0.0004966031685643632,
+      "loss": 1.7018,
+      "mean_token_accuracy": 0.5984292014539242,
       "step": 500
     },
     {
       "epoch": 0.013587233435464039,
-      "grad_norm": 0.4064154028892517,
-      "learning_rate": 0.0004932267181173402,
-      "loss": 1.7339,
-      "mean_token_accuracy": 0.59465124219656,
+      "grad_norm": 0.1678491234779358,
+      "learning_rate": 0.0004932063371287264,
+      "loss": 1.7027,
+      "mean_token_accuracy": 0.598892825126648,
       "step": 1000
     },
     {
       "epoch": 0.020380850153196056,
-      "grad_norm": 0.42659252882003784,
-      "learning_rate": 0.0004898298866817033,
-      "loss": 1.7231,
-      "mean_token_accuracy": 0.5960495759248734,
+      "grad_norm": 0.2502967119216919,
+      "learning_rate": 0.000489823093018832,
+      "loss": 1.6937,
+      "mean_token_accuracy": 0.5998013806045055,
       "step": 1500
     },
     {
       "epoch": 0.027174466870928077,
-      "grad_norm": 0.37063390016555786,
-      "learning_rate": 0.00048643305524606644,
-      "loss": 1.7228,
-      "mean_token_accuracy": 0.5968798612058163,
+      "grad_norm": 0.17444606125354767,
+      "learning_rate": 0.0004864262615831952,
+      "loss": 1.6963,
+      "mean_token_accuracy": 0.6011586433649063,
       "step": 2000
     },
     {
       "epoch": 0.033968083588660096,
-      "grad_norm": 0.4430277943611145,
+      "grad_norm": 0.20226487517356873,
       "learning_rate": 0.0004830430174733009,
-      "loss": 1.7177,
-      "mean_token_accuracy": 0.5959446367025375,
+      "loss": 1.6957,
+      "mean_token_accuracy": 0.5982117947936058,
       "step": 2500
     },
     {
       "epoch": 0.04076170030639211,
-      "grad_norm": 0.5757402181625366,
-      "learning_rate": 0.00047965297970053533,
-      "loss": 1.7291,
-      "mean_token_accuracy": 0.5950138145983219,
+      "grad_norm": 0.22139909863471985,
+      "learning_rate": 0.0004796461860376641,
+      "loss": 1.7079,
+      "mean_token_accuracy": 0.5977413670122623,
       "step": 3000
     },
     {
       "epoch": 0.04755531702412413,
-      "grad_norm": 0.465655654668808,
-      "learning_rate": 0.0004762561482648985,
-      "loss": 1.7146,
-      "mean_token_accuracy": 0.5979318037629128,
+      "grad_norm": 0.19102013111114502,
+      "learning_rate": 0.0004762629419277698,
+      "loss": 1.6945,
+      "mean_token_accuracy": 0.6003216138780118,
       "step": 3500
     },
     {
       "epoch": 0.054348933741856155,
-      "grad_norm": 0.3922131061553955,
-      "learning_rate": 0.00047285931682926167,
-      "loss": 1.7188,
-      "mean_token_accuracy": 0.593923301845789,
+      "grad_norm": 0.2559124529361725,
+      "learning_rate": 0.000472866110492133,
+      "loss": 1.7,
+      "mean_token_accuracy": 0.596915179759264,
       "step": 4000
     },
     {
       "epoch": 0.06114255045958817,
-      "grad_norm": 0.5041131377220154,
-      "learning_rate": 0.00046946248539362484,
-      "loss": 1.6899,
-      "mean_token_accuracy": 0.6019016440212727,
+      "grad_norm": 0.1964866667985916,
+      "learning_rate": 0.00046948286638223864,
+      "loss": 1.6714,
+      "mean_token_accuracy": 0.6040974244773388,
       "step": 4500
     },
     {
       "epoch": 0.06793616717732019,
-      "grad_norm": 0.3338882327079773,
-      "learning_rate": 0.00046607244762085926,
-      "loss": 1.7051,
-      "mean_token_accuracy": 0.5998869875967503,
+      "grad_norm": 0.15417207777500153,
+      "learning_rate": 0.00046608603494660187,
+      "loss": 1.688,
+      "mean_token_accuracy": 0.6016907621920109,
       "step": 5000
     },
     {
       "epoch": 0.07472978389505221,
-      "grad_norm": 0.4194291830062866,
-      "learning_rate": 0.00046267561618522243,
-      "loss": 1.7102,
-      "mean_token_accuracy": 0.5985791925489903,
+      "grad_norm": 0.20179523527622223,
+      "learning_rate": 0.00046270279083670754,
+      "loss": 1.6929,
+      "mean_token_accuracy": 0.600579642444849,
       "step": 5500
     },
     {
       "epoch": 0.08152340061278422,
-      "grad_norm": 0.42370814085006714,
-      "learning_rate": 0.0004592991657381994,
-      "loss": 1.7104,
-      "mean_token_accuracy": 0.5979952279627323,
+      "grad_norm": 0.22086994349956512,
+      "learning_rate": 0.00045930595940107065,
+      "loss": 1.694,
+      "mean_token_accuracy": 0.6007796683311463,
       "step": 6000
     },
     {
       "epoch": 0.08831701733051625,
-      "grad_norm": 0.43259018659591675,
-      "learning_rate": 0.0004559023343025626,
-      "loss": 1.7007,
-      "mean_token_accuracy": 0.6013973373174667,
+      "grad_norm": 0.16984662413597107,
+      "learning_rate": 0.00045592271529117643,
+      "loss": 1.6843,
+      "mean_token_accuracy": 0.6042516341209412,
       "step": 6500
     },
     {
       "epoch": 0.09511063404824827,
-      "grad_norm": 0.4605717957019806,
-      "learning_rate": 0.00045250550286692574,
-      "loss": 1.7022,
-      "mean_token_accuracy": 0.5996414051353931,
+      "grad_norm": 0.27594026923179626,
+      "learning_rate": 0.00045252588385553955,
+      "loss": 1.6872,
+      "mean_token_accuracy": 0.6015865850150586,
       "step": 7000
     },
     {
       "epoch": 0.10190425076598028,
-      "grad_norm": 0.4562588632106781,
-      "learning_rate": 0.0004491086714312889,
-      "loss": 1.7047,
-      "mean_token_accuracy": 0.5984801447093486,
+      "grad_norm": 0.14912690222263336,
+      "learning_rate": 0.00044914263974564527,
+      "loss": 1.6912,
+      "mean_token_accuracy": 0.6005175986289978,
       "step": 7500
     },
     {
       "epoch": 0.10869786748371231,
-      "grad_norm": 0.43784981966018677,
-      "learning_rate": 0.0004457254273213946,
-      "loss": 1.6963,
-      "mean_token_accuracy": 0.6002669661641121,
+      "grad_norm": 0.21813392639160156,
+      "learning_rate": 0.00044574580831000844,
+      "loss": 1.6821,
+      "mean_token_accuracy": 0.6029459582269192,
       "step": 8000
     },
     {
       "epoch": 0.11549148420144432,
-      "grad_norm": 0.4280397891998291,
-      "learning_rate": 0.0004423285958857578,
-      "loss": 1.7075,
-      "mean_token_accuracy": 0.5975034932494163,
+      "grad_norm": 0.21067260205745697,
+      "learning_rate": 0.00044236256420011416,
+      "loss": 1.6939,
+      "mean_token_accuracy": 0.5990548622012138,
       "step": 8500
     },
     {
       "epoch": 0.12228510091917634,
-      "grad_norm": 0.4308539628982544,
-      "learning_rate": 0.000438931764450121,
-      "loss": 1.6968,
-      "mean_token_accuracy": 0.599761339455843,
+      "grad_norm": 0.21087749302387238,
+      "learning_rate": 0.00043896573276447733,
+      "loss": 1.6846,
+      "mean_token_accuracy": 0.601339833676815,
       "step": 9000
     },
     {
       "epoch": 0.12907871763690837,
-      "grad_norm": 0.37971165776252747,
-      "learning_rate": 0.0004355349330144841,
-      "loss": 1.6893,
-      "mean_token_accuracy": 0.602813461214304,
+      "grad_norm": 0.18933086097240448,
+      "learning_rate": 0.000435582488654583,
+      "loss": 1.6774,
+      "mean_token_accuracy": 0.6039153957962989,
       "step": 9500
     },
     {
       "epoch": 0.13587233435464038,
-      "grad_norm": 0.42085403203964233,
-      "learning_rate": 0.0004321516889045898,
-      "loss": 1.6944,
-      "mean_token_accuracy": 0.5994773605763912,
+      "grad_norm": 0.22444137930870056,
+      "learning_rate": 0.00043218565721894617,
+      "loss": 1.6841,
+      "mean_token_accuracy": 0.6008755776882172,
       "step": 10000
     },
     {
       "epoch": 0.1426659510723724,
-      "grad_norm": 0.5180936455726624,
-      "learning_rate": 0.00042876165113182424,
-      "loss": 1.6939,
-      "mean_token_accuracy": 0.5987960757911205,
+      "grad_norm": 0.22333703935146332,
+      "learning_rate": 0.0004288024131090519,
+      "loss": 1.6833,
+      "mean_token_accuracy": 0.6002273961901665,
       "step": 10500
     },
     {
       "epoch": 0.14945956779010441,
-      "grad_norm": 0.49184420704841614,
-      "learning_rate": 0.0004253648196961874,
-      "loss": 1.6819,
-      "mean_token_accuracy": 0.6046352021694184,
+      "grad_norm": 0.29101186990737915,
+      "learning_rate": 0.000425405581673415,
+      "loss": 1.6717,
+      "mean_token_accuracy": 0.6063305280804634,
       "step": 11000
     },
     {
       "epoch": 0.15625318450783643,
-      "grad_norm": 0.4975549578666687,
-      "learning_rate": 0.0004219679882605506,
-      "loss": 1.6899,
-      "mean_token_accuracy": 0.6011166540384293,
+      "grad_norm": 0.24681076407432556,
+      "learning_rate": 0.0004220223375635208,
+      "loss": 1.681,
+      "mean_token_accuracy": 0.6020885335206986,
       "step": 11500
     },
     {
       "epoch": 0.16304680122556844,
-      "grad_norm": 0.583638072013855,
-      "learning_rate": 0.00041857115682491375,
-      "loss": 1.6809,
-      "mean_token_accuracy": 0.6029804965853691,
+      "grad_norm": 0.20586206018924713,
+      "learning_rate": 0.0004186255061278839,
+      "loss": 1.672,
+      "mean_token_accuracy": 0.6043462678194046,
       "step": 12000
     },
     {
       "epoch": 0.1698404179433005,
-      "grad_norm": 0.46068498492240906,
-      "learning_rate": 0.0004151947063778907,
-      "loss": 1.6924,
-      "mean_token_accuracy": 0.5984785186350345,
+      "grad_norm": 0.2876848578453064,
+      "learning_rate": 0.00041524226201798963,
+      "loss": 1.6826,
+      "mean_token_accuracy": 0.6002557762265205,
       "step": 12500
     },
     {
       "epoch": 0.1766340346610325,
-      "grad_norm": 0.4548553228378296,
-      "learning_rate": 0.00041179787494225384,
-      "loss": 1.7141,
-      "mean_token_accuracy": 0.5960147912800312,
+      "grad_norm": 0.22999879717826843,
+      "learning_rate": 0.0004118454305823528,
+      "loss": 1.7055,
+      "mean_token_accuracy": 0.5974675108492374,
       "step": 13000
     },
     {
       "epoch": 0.18342765137876452,
-      "grad_norm": 0.35259950160980225,
-      "learning_rate": 0.0004084078371694883,
-      "loss": 1.7108,
-      "mean_token_accuracy": 0.59765238276124,
+      "grad_norm": 0.19900381565093994,
+      "learning_rate": 0.00040846218647245847,
+      "loss": 1.7018,
+      "mean_token_accuracy": 0.5989092390239239,
       "step": 13500
     },
     {
       "epoch": 0.19022126809649653,
-      "grad_norm": 0.550769567489624,
-      "learning_rate": 0.0004050110057338515,
-      "loss": 1.6857,
-      "mean_token_accuracy": 0.6022021851241589,
+      "grad_norm": 0.2693159878253937,
+      "learning_rate": 0.0004050653550368217,
+      "loss": 1.6768,
+      "mean_token_accuracy": 0.6034152474999428,
       "step": 14000
     },
     {
       "epoch": 0.19701488481422855,
-      "grad_norm": 0.46585261821746826,
-      "learning_rate": 0.0004016141742982146,
-      "loss": 1.68,
-      "mean_token_accuracy": 0.6016397434175015,
+      "grad_norm": 0.21865518391132355,
+      "learning_rate": 0.00040168211092692736,
+      "loss": 1.6718,
+      "mean_token_accuracy": 0.6028618058264256,
       "step": 14500
     },
     {
       "epoch": 0.20380850153196056,
-      "grad_norm": 0.463501513004303,
-      "learning_rate": 0.00039821734286257777,
-      "loss": 1.6799,
-      "mean_token_accuracy": 0.6029948250949383,
+      "grad_norm": 0.2849615812301636,
+      "learning_rate": 0.00039828527949129053,
+      "loss": 1.6714,
+      "mean_token_accuracy": 0.6048314180970192,
       "step": 15000
     },
     {
       "epoch": 0.21060211824969258,
-      "grad_norm": 0.4906240701675415,
-      "learning_rate": 0.000394820511426941,
-      "loss": 1.699,
-      "mean_token_accuracy": 0.6011771605908871,
+      "grad_norm": 0.1779908388853073,
+      "learning_rate": 0.00039490203538139625,
+      "loss": 1.6911,
+      "mean_token_accuracy": 0.6023711397051811,
       "step": 15500
     },
     {
       "epoch": 0.21739573496742462,
-      "grad_norm": 0.49514085054397583,
-      "learning_rate": 0.0003914236799913041,
-      "loss": 1.6805,
-      "mean_token_accuracy": 0.6031656765639782,
+      "grad_norm": 0.23363584280014038,
+      "learning_rate": 0.00039150520394575937,
+      "loss": 1.6724,
+      "mean_token_accuracy": 0.6044722346067428,
       "step": 16000
     },
     {
       "epoch": 0.22418935168515663,
-      "grad_norm": 0.5545573830604553,
-      "learning_rate": 0.00038804043588140983,
-      "loss": 1.6914,
-      "mean_token_accuracy": 0.6014098982512951,
+      "grad_norm": 0.23164866864681244,
+      "learning_rate": 0.00038812195983586515,
+      "loss": 1.6821,
+      "mean_token_accuracy": 0.6025199710726737,
       "step": 16500
     },
     {
       "epoch": 0.23098296840288865,
-      "grad_norm": 0.4257826805114746,
-      "learning_rate": 0.00038465039810864425,
-      "loss": 1.6914,
-      "mean_token_accuracy": 0.6002418109178543,
+      "grad_norm": 0.2129431664943695,
+      "learning_rate": 0.00038472512840022826,
+      "loss": 1.6834,
+      "mean_token_accuracy": 0.6014123694598674,
       "step": 17000
     },
     {
       "epoch": 0.23777658512062066,
-      "grad_norm": 0.47851425409317017,
-      "learning_rate": 0.0003812535666730074,
-      "loss": 1.6871,
-      "mean_token_accuracy": 0.6003048142194748,
+      "grad_norm": 0.27956533432006836,
+      "learning_rate": 0.000381341884290334,
+      "loss": 1.6803,
+      "mean_token_accuracy": 0.6010346503853798,
       "step": 17500
     },
     {
       "epoch": 0.24457020183835268,
-      "grad_norm": 0.438859760761261,
-      "learning_rate": 0.0003778567352373706,
-      "loss": 1.6891,
-      "mean_token_accuracy": 0.5999817061126232,
+      "grad_norm": 0.3014739751815796,
+      "learning_rate": 0.00037794505285469716,
+      "loss": 1.6826,
+      "mean_token_accuracy": 0.6009049181342125,
       "step": 18000
     },
     {
       "epoch": 0.2513638185560847,
-      "grad_norm": 0.39294207096099854,
-      "learning_rate": 0.0003744599038017337,
-      "loss": 1.6881,
-      "mean_token_accuracy": 0.6006654644608498,
+      "grad_norm": 0.21182464063167572,
+      "learning_rate": 0.0003745618087448028,
+      "loss": 1.6804,
+      "mean_token_accuracy": 0.6019478496909142,
       "step": 18500
     },
     {
       "epoch": 0.25815743527381674,
-      "grad_norm": 0.40909263491630554,
-      "learning_rate": 0.0003710766596918395,
-      "loss": 1.6846,
-      "mean_token_accuracy": 0.6008345121741295,
+      "grad_norm": 0.22530542314052582,
+      "learning_rate": 0.00037116497730916605,
+      "loss": 1.6778,
+      "mean_token_accuracy": 0.6016871838569641,
       "step": 19000
     },
     {
       "epoch": 0.2649510519915487,
-      "grad_norm": 0.48700499534606934,
-      "learning_rate": 0.0003676798282562026,
-      "loss": 1.6855,
-      "mean_token_accuracy": 0.6010488954186439,
+      "grad_norm": 0.20813944935798645,
+      "learning_rate": 0.0003677817331992717,
+      "loss": 1.6779,
+      "mean_token_accuracy": 0.6021616536378861,
       "step": 19500
     },
     {
       "epoch": 0.27174466870928077,
-      "grad_norm": 0.5421476364135742,
-      "learning_rate": 0.0003642965841463083,
-      "loss": 1.6855,
-      "mean_token_accuracy": 0.599077152788639,
+      "grad_norm": 0.2850479185581207,
+      "learning_rate": 0.0003643849017636349,
+      "loss": 1.677,
+      "mean_token_accuracy": 0.600636290282011,
       "step": 20000
     },
     {
       "epoch": 0.27853828542701276,
-      "grad_norm": 0.7285926342010498,
-      "learning_rate": 0.0003608997527106715,
-      "loss": 1.6952,
-      "mean_token_accuracy": 0.5991786162257194,
+      "grad_norm": 0.22972407937049866,
+      "learning_rate": 0.0003610016576537406,
+      "loss": 1.687,
+      "mean_token_accuracy": 0.6000275067687034,
       "step": 20500
     },
     {
       "epoch": 0.2853319021447448,
-      "grad_norm": 0.4716951549053192,
-      "learning_rate": 0.0003575165086007772,
-      "loss": 1.6799,
-      "mean_token_accuracy": 0.6026499591767788,
+      "grad_norm": 0.2997701168060303,
+      "learning_rate": 0.0003576048262181037,
+      "loss": 1.6728,
+      "mean_token_accuracy": 0.6031140223443509,
       "step": 21000
     },
     {
       "epoch": 0.29212551886247684,
-      "grad_norm": 0.4939189851284027,
-      "learning_rate": 0.0003541196771651404,
-      "loss": 1.6927,
-      "mean_token_accuracy": 0.5991746371984482,
+      "grad_norm": 0.32064828276634216,
+      "learning_rate": 0.0003542215821082095,
+      "loss": 1.6856,
+      "mean_token_accuracy": 0.5996782946884632,
       "step": 21500
     },
     {
       "epoch": 0.29891913558020883,
-      "grad_norm": 0.3434264361858368,
-      "learning_rate": 0.00035073643305524605,
-      "loss": 1.6868,
-      "mean_token_accuracy": 0.6017061140835285,
+      "grad_norm": 0.1890692114830017,
+      "learning_rate": 0.0003508315443354439,
+      "loss": 1.6789,
+      "mean_token_accuracy": 0.603088031232357,
       "step": 22000
     },
     {
       "epoch": 0.30571275229794087,
-      "grad_norm": 0.4739142060279846,
-      "learning_rate": 0.0003473396016196093,
-      "loss": 1.6943,
-      "mean_token_accuracy": 0.5989596297442913,
+      "grad_norm": 0.277016282081604,
+      "learning_rate": 0.00034743471289980704,
+      "loss": 1.6862,
+      "mean_token_accuracy": 0.6002373751103878,
       "step": 22500
     },
     {
       "epoch": 0.31250636901567286,
-      "grad_norm": 0.6250849962234497,
-      "learning_rate": 0.00034395635750971495,
-      "loss": 1.6738,
-      "mean_token_accuracy": 0.6042404849231243,
+      "grad_norm": 0.3048676550388336,
+      "learning_rate": 0.00034403788146417026,
+      "loss": 1.6667,
+      "mean_token_accuracy": 0.6053032230436802,
       "step": 23000
     },
     {
       "epoch": 0.3192999857334049,
-      "grad_norm": 0.4180070459842682,
-      "learning_rate": 0.00034055952607407806,
-      "loss": 1.6758,
-      "mean_token_accuracy": 0.6041967616379261,
+      "grad_norm": 0.23588168621063232,
+      "learning_rate": 0.0003406410500285334,
+      "loss": 1.6687,
+      "mean_token_accuracy": 0.6044456725120544,
       "step": 23500
     },
     {
       "epoch": 0.3260936024511369,
-      "grad_norm": 0.4894815981388092,
-      "learning_rate": 0.00033717628196418384,
-      "loss": 1.6799,
-      "mean_token_accuracy": 0.6030513152778149,
+      "grad_norm": 0.34810811281204224,
+      "learning_rate": 0.0003372578059186391,
+      "loss": 1.6727,
+      "mean_token_accuracy": 0.603928690880537,
       "step": 24000
     },
     {
       "epoch": 0.33288721916886893,
-      "grad_norm": 0.4394868314266205,
-      "learning_rate": 0.00033377945052854695,
-      "loss": 1.6896,
-      "mean_token_accuracy": 0.5995272571742535,
+      "grad_norm": 0.2968311011791229,
+      "learning_rate": 0.00033386097448300227,
+      "loss": 1.6827,
+      "mean_token_accuracy": 0.599886593401432,
       "step": 24500
     },
     {
       "epoch": 0.339680835886601,
-      "grad_norm": 0.4560585021972656,
-      "learning_rate": 0.0003303962064186527,
-      "loss": 1.6761,
-      "mean_token_accuracy": 0.602882645457983,
+      "grad_norm": 0.2510404586791992,
+      "learning_rate": 0.00033047773037310794,
+      "loss": 1.6693,
+      "mean_token_accuracy": 0.6038261016011238,
       "step": 25000
     },
     {
       "epoch": 0.34647445260433296,
-      "grad_norm": 0.4372274875640869,
-      "learning_rate": 0.00032699937498301585,
-      "loss": 1.6729,
-      "mean_token_accuracy": 0.6026374279558658,
+      "grad_norm": 0.31068530678749084,
+      "learning_rate": 0.0003270808989374711,
+      "loss": 1.6647,
+      "mean_token_accuracy": 0.6035428291857242,
       "step": 25500
     },
     {
       "epoch": 0.353268069322065,
-      "grad_norm": 0.40903204679489136,
-      "learning_rate": 0.00032361613087312157,
-      "loss": 1.6732,
-      "mean_token_accuracy": 0.6029972824752331,
+      "grad_norm": 0.17706353962421417,
+      "learning_rate": 0.00032369765482757683,
+      "loss": 1.6655,
+      "mean_token_accuracy": 0.603890900760889,
       "step": 26000
     },
     {
       "epoch": 0.360061686039797,
-      "grad_norm": 0.4154110550880432,
-      "learning_rate": 0.00032021929943748474,
-      "loss": 1.6889,
-      "mean_token_accuracy": 0.6004891120493412,
+      "grad_norm": 0.29017922282218933,
+      "learning_rate": 0.00032030082339194,
+      "loss": 1.6817,
+      "mean_token_accuracy": 0.6018013699650765,
       "step": 26500
     },
     {
       "epoch": 0.36685530275752903,
-      "grad_norm": 0.47997039556503296,
-      "learning_rate": 0.0003168360553275904,
-      "loss": 1.6688,
-      "mean_token_accuracy": 0.6037946022450924,
+      "grad_norm": 0.23541323840618134,
+      "learning_rate": 0.00031691757928204573,
+      "loss": 1.6628,
+      "mean_token_accuracy": 0.6049926683902741,
       "step": 27000
     },
     {
       "epoch": 0.373648919475261,
-      "grad_norm": 0.4413166046142578,
-      "learning_rate": 0.00031343922389195363,
-      "loss": 1.6656,
-      "mean_token_accuracy": 0.6043658980131149,
+      "grad_norm": 0.2717786431312561,
+      "learning_rate": 0.0003135207478464089,
+      "loss": 1.6584,
+      "mean_token_accuracy": 0.6049346162080764,
       "step": 27500
     },
     {
       "epoch": 0.38044253619299306,
-      "grad_norm": 0.3751048147678375,
-      "learning_rate": 0.0003100559797820593,
-      "loss": 1.6813,
-      "mean_token_accuracy": 0.601385792285204,
+      "grad_norm": 0.20958386361598969,
+      "learning_rate": 0.00031013750373651457,
+      "loss": 1.6746,
+      "mean_token_accuracy": 0.6019823787510395,
       "step": 28000
     },
     {
       "epoch": 0.3872361529107251,
-      "grad_norm": 0.42594870924949646,
-      "learning_rate": 0.0003066591483464224,
-      "loss": 1.693,
-      "mean_token_accuracy": 0.601170395553112,
+      "grad_norm": 0.26794660091400146,
+      "learning_rate": 0.0003067406723008778,
+      "loss": 1.6869,
+      "mean_token_accuracy": 0.6023132638335228,
       "step": 28500
     },
     {
       "epoch": 0.3940297696284571,
-      "grad_norm": 0.5058168172836304,
-      "learning_rate": 0.0003032759042365282,
-      "loss": 1.6761,
-      "mean_token_accuracy": 0.6033368219733238,
+      "grad_norm": 0.23140773177146912,
+      "learning_rate": 0.00030335742819098346,
+      "loss": 1.6687,
+      "mean_token_accuracy": 0.6039072838723659,
       "step": 29000
     },
     {
       "epoch": 0.40082338634618914,
-      "grad_norm": 0.5077099800109863,
-      "learning_rate": 0.0002998790728008913,
-      "loss": 1.6726,
-      "mean_token_accuracy": 0.6028481238484382,
+      "grad_norm": 0.2647615969181061,
+      "learning_rate": 0.0002999605967553466,
+      "loss": 1.6657,
+      "mean_token_accuracy": 0.6038105982542038,
       "step": 29500
     },
     {
       "epoch": 0.4076170030639211,
-      "grad_norm": 0.4526306986808777,
-      "learning_rate": 0.00029649582869099704,
-      "loss": 1.6955,
-      "mean_token_accuracy": 0.6015688245892524,
+      "grad_norm": 0.3417193591594696,
+      "learning_rate": 0.0002965773526454523,
+      "loss": 1.6886,
+      "mean_token_accuracy": 0.6023576308786869,
       "step": 30000
     },
     {
       "epoch": 0.41441061978165317,
-      "grad_norm": 0.7025688886642456,
-      "learning_rate": 0.0002930989972553602,
-      "loss": 1.6804,
-      "mean_token_accuracy": 0.6013101308941841,
+      "grad_norm": 0.30876556038856506,
+      "learning_rate": 0.00029318052120981547,
+      "loss": 1.6732,
+      "mean_token_accuracy": 0.602056191176176,
       "step": 30500
     },
     {
       "epoch": 0.42120423649938515,
-      "grad_norm": 0.37353208661079407,
-      "learning_rate": 0.0002897089594825946,
-      "loss": 1.6616,
-      "mean_token_accuracy": 0.6058516998887062,
+      "grad_norm": 0.22551144659519196,
+      "learning_rate": 0.0002897972770999212,
+      "loss": 1.6547,
+      "mean_token_accuracy": 0.6065524759292602,
       "step": 31000
     },
     {
       "epoch": 0.4279978532171172,
-      "grad_norm": 0.40674665570259094,
-      "learning_rate": 0.0002863121280469578,
-      "loss": 1.6714,
-      "mean_token_accuracy": 0.601816111266613,
+      "grad_norm": 0.2366376370191574,
+      "learning_rate": 0.00028640044566428436,
+      "loss": 1.6664,
+      "mean_token_accuracy": 0.6022460307478905,
       "step": 31500
     },
     {
       "epoch": 0.43479146993484924,
-      "grad_norm": 0.49832838773727417,
-      "learning_rate": 0.00028291529661132096,
-      "loss": 1.6941,
-      "mean_token_accuracy": 0.5991431129276752,
+      "grad_norm": 0.27979329228401184,
+      "learning_rate": 0.0002830172015543901,
+      "loss": 1.6878,
+      "mean_token_accuracy": 0.5996840268671513,
       "step": 32000
     },
     {
       "epoch": 0.4415850866525812,
-      "grad_norm": 0.47476768493652344,
-      "learning_rate": 0.00027951846517568414,
-      "loss": 1.6857,
-      "mean_token_accuracy": 0.5986996866464614,
+      "grad_norm": 0.24599482119083405,
+      "learning_rate": 0.00027962037011875326,
+      "loss": 1.6795,
+      "mean_token_accuracy": 0.5997087613940238,
       "step": 32500
     },
     {
       "epoch": 0.44837870337031327,
-      "grad_norm": 0.4772721230983734,
-      "learning_rate": 0.0002761352210657898,
-      "loss": 1.6699,
-      "mean_token_accuracy": 0.6043048515915871,
+      "grad_norm": 0.23995830118656158,
+      "learning_rate": 0.0002762371260088589,
+      "loss": 1.6637,
+      "mean_token_accuracy": 0.6052455662488937,
       "step": 33000
     },
     {
       "epoch": 0.45517232008804526,
-      "grad_norm": 0.43461740016937256,
-      "learning_rate": 0.00027273838963015303,
-      "loss": 1.6807,
-      "mean_token_accuracy": 0.6015589760541916,
+      "grad_norm": 0.2345362901687622,
+      "learning_rate": 0.00027284029457322215,
+      "loss": 1.6749,
+      "mean_token_accuracy": 0.602227875560522,
       "step": 33500
     },
     {
       "epoch": 0.4619659368057773,
-      "grad_norm": 0.49901214241981506,
-      "learning_rate": 0.0002693551455202587,
-      "loss": 1.6693,
-      "mean_token_accuracy": 0.603463482439518,
+      "grad_norm": 0.25029075145721436,
+      "learning_rate": 0.0002694570504633278,
+      "loss": 1.6625,
+      "mean_token_accuracy": 0.6042112992107869,
       "step": 34000
     },
     {
       "epoch": 0.4687595535235093,
-      "grad_norm": 0.5628573298454285,
-      "learning_rate": 0.00026595831408462187,
-      "loss": 1.661,
-      "mean_token_accuracy": 0.6067080739736557,
+      "grad_norm": 0.24419337511062622,
+      "learning_rate": 0.00026606021902769093,
+      "loss": 1.6553,
+      "mean_token_accuracy": 0.6069871078431607,
       "step": 34500
     },
     {
       "epoch": 0.47555317024124133,
-      "grad_norm": 0.517264187335968,
-      "learning_rate": 0.0002625750699747276,
-      "loss": 1.6679,
-      "mean_token_accuracy": 0.6032546425163746,
+      "grad_norm": 0.23328031599521637,
+      "learning_rate": 0.0002626769749177967,
+      "loss": 1.6618,
+      "mean_token_accuracy": 0.6040745720565319,
       "step": 35000
     },
     {
       "epoch": 0.4823467869589734,
-      "grad_norm": 0.45866549015045166,
-      "learning_rate": 0.0002591782385390907,
-      "loss": 1.6614,
-      "mean_token_accuracy": 0.6066102555394173,
+      "grad_norm": 0.23471958935260773,
+      "learning_rate": 0.0002592801434821598,
+      "loss": 1.6552,
+      "mean_token_accuracy": 0.607022188782692,
       "step": 35500
     },
     {
       "epoch": 0.48914040367670536,
-      "grad_norm": 0.4672979414463043,
-      "learning_rate": 0.0002557949944291965,
-      "loss": 1.6736,
-      "mean_token_accuracy": 0.6030917878448964,
+      "grad_norm": 0.2557450830936432,
+      "learning_rate": 0.00025589689937226555,
+      "loss": 1.6682,
+      "mean_token_accuracy": 0.6036945071816444,
       "step": 36000
     },
     {
       "epoch": 0.4959340203944374,
-      "grad_norm": 0.43743544816970825,
-      "learning_rate": 0.00025240495665643085,
-      "loss": 1.6652,
-      "mean_token_accuracy": 0.6046005863845348,
+      "grad_norm": 0.32713398337364197,
+      "learning_rate": 0.0002525000679366287,
+      "loss": 1.6588,
+      "mean_token_accuracy": 0.6052843576967716,
       "step": 36500
     },
     {
       "epoch": 0.5027276371121694,
-      "grad_norm": 0.4721488058567047,
-      "learning_rate": 0.0002490081252207941,
-      "loss": 1.6795,
-      "mean_token_accuracy": 0.6025200285613537,
+      "grad_norm": 0.26545101404190063,
+      "learning_rate": 0.00024911682382673444,
+      "loss": 1.6735,
+      "mean_token_accuracy": 0.6035105584859848,
       "step": 37000
     },
     {
       "epoch": 0.5095212538299014,
-      "grad_norm": 0.44578176736831665,
-      "learning_rate": 0.0002456112937851572,
-      "loss": 1.6703,
-      "mean_token_accuracy": 0.6039398565590381,
+      "grad_norm": 0.21667255461215973,
+      "learning_rate": 0.0002457199923910976,
+      "loss": 1.6641,
+      "mean_token_accuracy": 0.6042561739683151,
       "step": 37500
     },
     {
       "epoch": 0.5163148705476335,
-      "grad_norm": 0.45349687337875366,
-      "learning_rate": 0.00024221446234952039,
-      "loss": 1.6655,
-      "mean_token_accuracy": 0.6041326204538345,
+      "grad_norm": 0.24779720604419708,
+      "learning_rate": 0.0002423367482812033,
+      "loss": 1.6584,
+      "mean_token_accuracy": 0.6050419315993786,
       "step": 38000
     },
     {
       "epoch": 0.5231084872653654,
-      "grad_norm": 0.3655533194541931,
-      "learning_rate": 0.00023883121823962608,
-      "loss": 1.676,
-      "mean_token_accuracy": 0.6035989390015603,
+      "grad_norm": 0.26215288043022156,
+      "learning_rate": 0.00023893991684556645,
+      "loss": 1.6711,
+      "mean_token_accuracy": 0.6044524771273136,
       "step": 38500
     },
     {
       "epoch": 0.5299021039830974,
-      "grad_norm": 0.5141535401344299,
-      "learning_rate": 0.00023543438680398923,
-      "loss": 1.6784,
-      "mean_token_accuracy": 0.6029777365326882,
+      "grad_norm": 0.31740403175354004,
+      "learning_rate": 0.00023555667273567217,
+      "loss": 1.6725,
+      "mean_token_accuracy": 0.6041430326402187,
       "step": 39000
     },
     {
       "epoch": 0.5366957207008295,
-      "grad_norm": 0.3538263142108917,
-      "learning_rate": 0.00023205114269409495,
-      "loss": 1.666,
-      "mean_token_accuracy": 0.6050683797895908,
+      "grad_norm": 0.24346832931041718,
+      "learning_rate": 0.00023215984130003534,
+      "loss": 1.6597,
+      "mean_token_accuracy": 0.6052628398537636,
       "step": 39500
     },
     {
       "epoch": 0.5434893374185615,
-      "grad_norm": 0.4689236283302307,
-      "learning_rate": 0.00022865431125845812,
-      "loss": 1.6608,
-      "mean_token_accuracy": 0.6069498200118542,
+      "grad_norm": 0.2325129359960556,
+      "learning_rate": 0.00022877659719014104,
+      "loss": 1.6555,
+      "mean_token_accuracy": 0.6073581319153308,
       "step": 40000
     },
     {
       "epoch": 0.5502829541362936,
-      "grad_norm": 0.4244871735572815,
-      "learning_rate": 0.00022527106714856381,
-      "loss": 1.6701,
-      "mean_token_accuracy": 0.6025859090983867,
+      "grad_norm": 0.23105812072753906,
+      "learning_rate": 0.00022537976575450418,
+      "loss": 1.6656,
+      "mean_token_accuracy": 0.6033125531077385,
       "step": 40500
     },
     {
       "epoch": 0.5570765708540255,
-      "grad_norm": 0.49384674429893494,
-      "learning_rate": 0.00022187423571292698,
-      "loss": 1.6827,
-      "mean_token_accuracy": 0.6007110984623432,
+      "grad_norm": 0.2570987343788147,
+      "learning_rate": 0.0002219965216446099,
+      "loss": 1.6783,
+      "mean_token_accuracy": 0.6010972237884998,
       "step": 41000
     },
     {
       "epoch": 0.5638701875717576,
-      "grad_norm": 0.5254752039909363,
-      "learning_rate": 0.0002184909916030327,
-      "loss": 1.661,
-      "mean_token_accuracy": 0.6038983342051506,
+      "grad_norm": 0.2975606620311737,
+      "learning_rate": 0.00021859969020897308,
+      "loss": 1.6558,
+      "mean_token_accuracy": 0.6043787428438664,
       "step": 41500
     },
     {
       "epoch": 0.5706638042894896,
-      "grad_norm": 0.5031416416168213,
-      "learning_rate": 0.00021509416016739585,
-      "loss": 1.667,
-      "mean_token_accuracy": 0.6047138755917549,
+      "grad_norm": 0.2746988534927368,
+      "learning_rate": 0.00021521644609907877,
+      "loss": 1.6622,
+      "mean_token_accuracy": 0.6047704361975194,
       "step": 42000
     },
     {
       "epoch": 0.5774574210072216,
-      "grad_norm": 0.5616060495376587,
-      "learning_rate": 0.00021171091605750157,
-      "loss": 1.6731,
-      "mean_token_accuracy": 0.6030472720265388,
+      "grad_norm": 0.2990375757217407,
+      "learning_rate": 0.00021181961466344194,
+      "loss": 1.6675,
+      "mean_token_accuracy": 0.6030395071804523,
       "step": 42500
     },
     {
       "epoch": 0.5842510377249537,
-      "grad_norm": 0.6484994292259216,
-      "learning_rate": 0.00020831408462186474,
-      "loss": 1.6687,
-      "mean_token_accuracy": 0.6059929049313069,
+      "grad_norm": 0.24443228542804718,
+      "learning_rate": 0.00020843637055354767,
+      "loss": 1.6633,
+      "mean_token_accuracy": 0.606367846518755,
       "step": 43000
     },
     {
       "epoch": 0.5910446544426856,
-      "grad_norm": 0.6379068493843079,
-      "learning_rate": 0.00020493084051197044,
-      "loss": 1.6919,
-      "mean_token_accuracy": 0.5994226909875869,
+      "grad_norm": 0.27472856640815735,
+      "learning_rate": 0.0002050395391179108,
+      "loss": 1.6881,
+      "mean_token_accuracy": 0.5997895430922509,
       "step": 43500
     },
     {
       "epoch": 0.5978382711604177,
-      "grad_norm": 0.5151463150978088,
-      "learning_rate": 0.0002015340090763336,
-      "loss": 1.6743,
-      "mean_token_accuracy": 0.6019911893308163,
+      "grad_norm": 0.2505660057067871,
+      "learning_rate": 0.00020165629500801653,
+      "loss": 1.6696,
+      "mean_token_accuracy": 0.601973987042904,
       "step": 44000
     },
     {
       "epoch": 0.6046318878781497,
-      "grad_norm": 0.4630017876625061,
-      "learning_rate": 0.0001981507649664393,
-      "loss": 1.6683,
-      "mean_token_accuracy": 0.6034873478114605,
+      "grad_norm": 0.22626104950904846,
+      "learning_rate": 0.0001982594635723797,
+      "loss": 1.6635,
+      "mean_token_accuracy": 0.6038794709444046,
       "step": 44500
     },
     {
       "epoch": 0.6114255045958817,
-      "grad_norm": 0.5501480102539062,
-      "learning_rate": 0.00019475393353080248,
-      "loss": 1.6675,
-      "mean_token_accuracy": 0.60292616507411,
+      "grad_norm": 0.2626524567604065,
+      "learning_rate": 0.0001948762194624854,
+      "loss": 1.6632,
+      "mean_token_accuracy": 0.6037038467228413,
       "step": 45000
     },
     {
       "epoch": 0.6182191213136137,
-      "grad_norm": 0.39942827820777893,
-      "learning_rate": 0.00019137068942090817,
-      "loss": 1.6877,
-      "mean_token_accuracy": 0.6008874903023242,
+      "grad_norm": 0.1837160289287567,
+      "learning_rate": 0.00019147938802684857,
+      "loss": 1.6835,
+      "mean_token_accuracy": 0.6012392081618309,
       "step": 45500
     },
     {
       "epoch": 0.6250127380313457,
-      "grad_norm": 0.4541276693344116,
-      "learning_rate": 0.00018797385798527134,
-      "loss": 1.6633,
-      "mean_token_accuracy": 0.6041645866334439,
+      "grad_norm": 0.47885841131210327,
+      "learning_rate": 0.00018809614391695426,
+      "loss": 1.6586,
+      "mean_token_accuracy": 0.6045624650716782,
       "step": 46000
     },
     {
       "epoch": 0.6318063547490778,
-      "grad_norm": 0.41248592734336853,
-      "learning_rate": 0.00018459061387537704,
-      "loss": 1.6802,
-      "mean_token_accuracy": 0.6013916251957416,
+      "grad_norm": 0.3030783236026764,
+      "learning_rate": 0.00018469931248131743,
+      "loss": 1.6759,
+      "mean_token_accuracy": 0.6016956052780151,
       "step": 46500
     },
     {
       "epoch": 0.6385999714668098,
-      "grad_norm": 0.5644799470901489,
-      "learning_rate": 0.0001811937824397402,
-      "loss": 1.6608,
-      "mean_token_accuracy": 0.6057685135900974,
+      "grad_norm": 0.2804652750492096,
+      "learning_rate": 0.00018131606837142313,
+      "loss": 1.6562,
+      "mean_token_accuracy": 0.606370843410492,
       "step": 47000
     },
     {
       "epoch": 0.6453935881845418,
-      "grad_norm": 0.4525160789489746,
-      "learning_rate": 0.00017781053832984593,
-      "loss": 1.6779,
-      "mean_token_accuracy": 0.6037380565404892,
+      "grad_norm": 0.26346665620803833,
+      "learning_rate": 0.0001779192369357863,
+      "loss": 1.6732,
+      "mean_token_accuracy": 0.6036491243541241,
       "step": 47500
     },
     {
       "epoch": 0.6521872049022738,
-      "grad_norm": 0.516008198261261,
-      "learning_rate": 0.00017441370689420907,
-      "loss": 1.6733,
-      "mean_token_accuracy": 0.6011509810388088,
+      "grad_norm": 0.26133692264556885,
+      "learning_rate": 0.000174535992825892,
+      "loss": 1.6683,
+      "mean_token_accuracy": 0.6019688593745232,
       "step": 48000
     },
     {
       "epoch": 0.6589808216200058,
-      "grad_norm": 0.5842276811599731,
-      "learning_rate": 0.0001710304627843148,
-      "loss": 1.6691,
-      "mean_token_accuracy": 0.6040294868052005,
+      "grad_norm": 0.3017692565917969,
+      "learning_rate": 0.00017113916139025517,
+      "loss": 1.665,
+      "mean_token_accuracy": 0.6044230219423771,
       "step": 48500
     },
     {
       "epoch": 0.6657744383377379,
-      "grad_norm": 0.5395560264587402,
-      "learning_rate": 0.00016763363134867797,
-      "loss": 1.6584,
-      "mean_token_accuracy": 0.6059439519941807,
+      "grad_norm": 0.24953605234622955,
+      "learning_rate": 0.0001677559172803609,
+      "loss": 1.6546,
+      "mean_token_accuracy": 0.606383463203907,
       "step": 49000
     },
     {
       "epoch": 0.6725680550554699,
-      "grad_norm": 0.4034149646759033,
-      "learning_rate": 0.00016425038723878366,
-      "loss": 1.6763,
-      "mean_token_accuracy": 0.6028572465479374,
+      "grad_norm": 0.2231573462486267,
+      "learning_rate": 0.00016435908584472403,
+      "loss": 1.6728,
+      "mean_token_accuracy": 0.6031719662547111,
       "step": 49500
     },
     {
       "epoch": 0.679361671773202,
-      "grad_norm": 0.4670972526073456,
-      "learning_rate": 0.00016085355580314683,
-      "loss": 1.6737,
-      "mean_token_accuracy": 0.6027044896781445,
+      "grad_norm": 0.25107043981552124,
+      "learning_rate": 0.00016097584173482976,
+      "loss": 1.6695,
+      "mean_token_accuracy": 0.6031585976481437,
       "step": 50000
     },
     {
       "epoch": 0.6861552884909339,
-      "grad_norm": 0.4919074475765228,
-      "learning_rate": 0.00015747031169325256,
-      "loss": 1.6724,
-      "mean_token_accuracy": 0.6025184873044491,
+      "grad_norm": 0.2994299829006195,
+      "learning_rate": 0.00015757901029919293,
+      "loss": 1.6695,
+      "mean_token_accuracy": 0.6030965386629105,
       "step": 50500
     },
     {
       "epoch": 0.6929489052086659,
-      "grad_norm": 0.5459266304969788,
-      "learning_rate": 0.0001540734802576157,
-      "loss": 1.6712,
-      "mean_token_accuracy": 0.601843435794115,
+      "grad_norm": 0.3250620663166046,
+      "learning_rate": 0.00015419576618929862,
+      "loss": 1.667,
+      "mean_token_accuracy": 0.6021432795226574,
       "step": 51000
     },
     {
       "epoch": 0.699742521926398,
-      "grad_norm": 0.5790686011314392,
-      "learning_rate": 0.0001506902361477214,
-      "loss": 1.6766,
-      "mean_token_accuracy": 0.6018453196883201,
+      "grad_norm": 0.29060378670692444,
+      "learning_rate": 0.0001507989347536618,
+      "loss": 1.6733,
+      "mean_token_accuracy": 0.6021323328912258,
       "step": 51500
     },
     {
       "epoch": 0.70653613864413,
-      "grad_norm": 0.6184782981872559,
-      "learning_rate": 0.00014729340471208457,
-      "loss": 1.6605,
-      "mean_token_accuracy": 0.6039663677811623,
+      "grad_norm": 0.29853805899620056,
+      "learning_rate": 0.00014741569064376751,
+      "loss": 1.6562,
+      "mean_token_accuracy": 0.6042688992321491,
       "step": 52000
     },
     {
       "epoch": 0.7133297553618619,
-      "grad_norm": 0.46344733238220215,
-      "learning_rate": 0.0001439101606021903,
-      "loss": 1.6668,
-      "mean_token_accuracy": 0.6035030370950699,
+      "grad_norm": 0.2511996328830719,
+      "learning_rate": 0.00014401885920813066,
+      "loss": 1.6631,
+      "mean_token_accuracy": 0.6040007803142071,
       "step": 52500
     },
     {
       "epoch": 0.720123372079594,
-      "grad_norm": 0.4513992965221405,
-      "learning_rate": 0.00014051332916655343,
-      "loss": 1.681,
-      "mean_token_accuracy": 0.6013197962939739,
+      "grad_norm": 0.24322637915611267,
+      "learning_rate": 0.00014063561509823635,
+      "loss": 1.6778,
+      "mean_token_accuracy": 0.6016692678034306,
       "step": 53000
     },
     {
       "epoch": 0.726916988797326,
-      "grad_norm": 0.6055275797843933,
-      "learning_rate": 0.00013713008505665913,
-      "loss": 1.6645,
-      "mean_token_accuracy": 0.6029207228422165,
+      "grad_norm": 0.36926066875457764,
+      "learning_rate": 0.00013723878366259952,
+      "loss": 1.6606,
+      "mean_token_accuracy": 0.6029434266984462,
       "step": 53500
     },
     {
       "epoch": 0.7337106055150581,
-      "grad_norm": 0.7768318057060242,
-      "learning_rate": 0.00013373325362102232,
-      "loss": 1.6723,
-      "mean_token_accuracy": 0.6039300902187824,
+      "grad_norm": 0.37792739272117615,
+      "learning_rate": 0.00013385553955270525,
+      "loss": 1.6697,
+      "mean_token_accuracy": 0.60406327226758,
       "step": 54000
     },
     {
       "epoch": 0.7405042222327901,
-      "grad_norm": 0.4635748863220215,
-      "learning_rate": 0.00013035000951112802,
-      "loss": 1.6755,
-      "mean_token_accuracy": 0.6006274480819702,
+      "grad_norm": 0.219089537858963,
+      "learning_rate": 0.0001304587081170684,
+      "loss": 1.6728,
+      "mean_token_accuracy": 0.6011269577145576,
       "step": 54500
     },
     {
       "epoch": 0.747297838950522,
-      "grad_norm": 0.45187345147132874,
-      "learning_rate": 0.00012695317807549116,
-      "loss": 1.6575,
-      "mean_token_accuracy": 0.6050706767737866,
+      "grad_norm": 0.2470192015171051,
+      "learning_rate": 0.0001270822576700454,
+      "loss": 1.6546,
+      "mean_token_accuracy": 0.605621423304081,
       "step": 55000
     },
     {
       "epoch": 0.7540914556682541,
-      "grad_norm": 0.4425111413002014,
-      "learning_rate": 0.0001235699339655969,
-      "loss": 1.6781,
-      "mean_token_accuracy": 0.6030424599051476,
+      "grad_norm": 0.22514095902442932,
+      "learning_rate": 0.00012368542623440856,
+      "loss": 1.6754,
+      "mean_token_accuracy": 0.6035324241816997,
       "step": 55500
     },
     {
       "epoch": 0.7608850723859861,
-      "grad_norm": 0.5505604147911072,
-      "learning_rate": 0.00012017310252996006,
-      "loss": 1.6554,
-      "mean_token_accuracy": 0.6084197340607643,
+      "grad_norm": 0.33952444791793823,
+      "learning_rate": 0.0001202885947987717,
+      "loss": 1.6533,
+      "mean_token_accuracy": 0.6087615159451961,
       "step": 56000
     },
     {
       "epoch": 0.7676786891037182,
-      "grad_norm": 0.5301482081413269,
-      "learning_rate": 0.00011678985842006575,
-      "loss": 1.6625,
-      "mean_token_accuracy": 0.602829116821289,
+      "grad_norm": 0.2792752683162689,
+      "learning_rate": 0.00011689176336313487,
+      "loss": 1.6598,
+      "mean_token_accuracy": 0.6029302821159362,
       "step": 56500
     },
     {
       "epoch": 0.7744723058214502,
-      "grad_norm": 0.5754754543304443,
-      "learning_rate": 0.00011339302698442892,
-      "loss": 1.6518,
-      "mean_token_accuracy": 0.6081080415248871,
+      "grad_norm": 0.3443421423435211,
+      "learning_rate": 0.00011350851925324058,
+      "loss": 1.6491,
+      "mean_token_accuracy": 0.6083803434073926,
       "step": 57000
     },
     {
       "epoch": 0.7812659225391821,
-      "grad_norm": 0.4835888743400574,
-      "learning_rate": 0.00011000978287453465,
-      "loss": 1.679,
-      "mean_token_accuracy": 0.6019827630519867,
+      "grad_norm": 0.2182215303182602,
+      "learning_rate": 0.00011011168781760374,
+      "loss": 1.6765,
+      "mean_token_accuracy": 0.6022524746358394,
       "step": 57500
     },
     {
       "epoch": 0.7880595392569142,
-      "grad_norm": 0.5146164894104004,
-      "learning_rate": 0.00010661295143889779,
-      "loss": 1.6735,
-      "mean_token_accuracy": 0.6020470056533813,
+      "grad_norm": 0.3579166531562805,
+      "learning_rate": 0.00010672165004483817,
+      "loss": 1.6715,
+      "mean_token_accuracy": 0.6020486257970333,
       "step": 58000
     },
     {
       "epoch": 0.7948531559746462,
-      "grad_norm": 0.5473788976669312,
-      "learning_rate": 0.00010322970732900351,
-      "loss": 1.6545,
-      "mean_token_accuracy": 0.6059571769535541,
+      "grad_norm": 0.2901332974433899,
+      "learning_rate": 0.00010332481860920134,
+      "loss": 1.6524,
+      "mean_token_accuracy": 0.6058744288384914,
       "step": 58500
     },
     {
       "epoch": 0.8016467726923783,
-      "grad_norm": 0.4318866729736328,
-      "learning_rate": 9.983287589336668e-05,
-      "loss": 1.6795,
-      "mean_token_accuracy": 0.6011920778155326,
+      "grad_norm": 0.24484753608703613,
+      "learning_rate": 9.992798717356451e-05,
+      "loss": 1.6776,
+      "mean_token_accuracy": 0.6012994230091572,
       "step": 59000
     },
     {
       "epoch": 0.8084403894101102,
-      "grad_norm": 0.5313432216644287,
-      "learning_rate": 9.644963178347238e-05,
-      "loss": 1.6498,
-      "mean_token_accuracy": 0.6086946985721589,
+      "grad_norm": 0.23479193449020386,
+      "learning_rate": 9.653115573792766e-05,
+      "loss": 1.6469,
+      "mean_token_accuracy": 0.6090488181114196,
       "step": 59500
     },
     {
       "epoch": 0.8152340061278422,
-      "grad_norm": 0.5233883261680603,
-      "learning_rate": 9.305280034783555e-05,
-      "loss": 1.663,
-      "mean_token_accuracy": 0.6036414765715599,
+      "grad_norm": 0.4180492162704468,
+      "learning_rate": 9.314791162803338e-05,
+      "loss": 1.6614,
+      "mean_token_accuracy": 0.6036504367887974,
       "step": 60000
     },
     {
       "epoch": 0.8220276228455743,
-      "grad_norm": 0.5144417881965637,
-      "learning_rate": 8.966955623794124e-05,
-      "loss": 1.6813,
-      "mean_token_accuracy": 0.6005311135351658,
+      "grad_norm": 0.3226146101951599,
+      "learning_rate": 8.975108019239653e-05,
+      "loss": 1.679,
+      "mean_token_accuracy": 0.6006873389482498,
       "step": 60500
     },
     {
       "epoch": 0.8288212395633063,
-      "grad_norm": 0.563146710395813,
-      "learning_rate": 8.627272480230441e-05,
-      "loss": 1.6721,
-      "mean_token_accuracy": 0.6010738264620304,
+      "grad_norm": 0.2912672460079193,
+      "learning_rate": 8.636783608250224e-05,
+      "loss": 1.6699,
+      "mean_token_accuracy": 0.6015279042720795,
       "step": 61000
     },
     {
       "epoch": 0.8356148562810384,
-      "grad_norm": 0.48985061049461365,
-      "learning_rate": 8.288948069241012e-05,
-      "loss": 1.6559,
-      "mean_token_accuracy": 0.6040105725228786,
+      "grad_norm": 0.22919562458992004,
+      "learning_rate": 8.297100464686541e-05,
+      "loss": 1.6547,
+      "mean_token_accuracy": 0.6040539441406727,
       "step": 61500
     },
     {
       "epoch": 0.8424084729987703,
-      "grad_norm": 0.3702440559864044,
-      "learning_rate": 7.949264925677328e-05,
-      "loss": 1.6629,
-      "mean_token_accuracy": 0.604276294618845,
+      "grad_norm": 0.18364617228507996,
+      "learning_rate": 7.958776053697111e-05,
+      "loss": 1.6617,
+      "mean_token_accuracy": 0.6046125743985176,
       "step": 62000
     },
     {
       "epoch": 0.8492020897165024,
-      "grad_norm": 0.4425947368144989,
-      "learning_rate": 7.610940514687899e-05,
-      "loss": 1.6484,
-      "mean_token_accuracy": 0.6068709415197372,
+      "grad_norm": 0.23012420535087585,
+      "learning_rate": 7.619092910133428e-05,
+      "loss": 1.6468,
+      "mean_token_accuracy": 0.6068955557644368,
       "step": 62500
     },
     {
       "epoch": 0.8559957064342344,
-      "grad_norm": 0.4522407352924347,
-      "learning_rate": 7.271257371124215e-05,
-      "loss": 1.6567,
-      "mean_token_accuracy": 0.6042509483098983,
+      "grad_norm": 0.36280351877212524,
+      "learning_rate": 7.280768499143999e-05,
+      "loss": 1.6555,
+      "mean_token_accuracy": 0.6039170788824558,
       "step": 63000
     },
     {
       "epoch": 0.8627893231519664,
-      "grad_norm": 0.4882795512676239,
-      "learning_rate": 6.932932960134787e-05,
-      "loss": 1.6548,
-      "mean_token_accuracy": 0.60615131264925,
+      "grad_norm": 0.24410812556743622,
+      "learning_rate": 6.941085355580315e-05,
+      "loss": 1.6531,
+      "mean_token_accuracy": 0.606347491145134,
       "step": 63500
     },
     {
       "epoch": 0.8695829398696985,
-      "grad_norm": 0.5515041351318359,
-      "learning_rate": 6.593249816571103e-05,
-      "loss": 1.663,
-      "mean_token_accuracy": 0.6043142848014832,
+      "grad_norm": 0.2840823233127594,
+      "learning_rate": 6.602760944590886e-05,
+      "loss": 1.6622,
+      "mean_token_accuracy": 0.6042196550667286,
       "step": 64000
     },
     {
       "epoch": 0.8763765565874304,
-      "grad_norm": 0.5676508545875549,
-      "learning_rate": 6.254925405581674e-05,
-      "loss": 1.6715,
-      "mean_token_accuracy": 0.6038499280512333,
+      "grad_norm": 0.2685336768627167,
+      "learning_rate": 6.263757167314329e-05,
+      "loss": 1.6711,
+      "mean_token_accuracy": 0.6039770551323891,
       "step": 64500
     },
     {
       "epoch": 0.8831701733051625,
-      "grad_norm": 0.6217424273490906,
-      "learning_rate": 5.91524226201799e-05,
-      "loss": 1.6616,
-      "mean_token_accuracy": 0.6056457755565643,
+      "grad_norm": 0.3238789737224579,
+      "learning_rate": 5.924074023750645e-05,
+      "loss": 1.6604,
+      "mean_token_accuracy": 0.6058327569365501,
       "step": 65000
     },
     {
       "epoch": 0.8899637900228945,
-      "grad_norm": 0.4277644157409668,
-      "learning_rate": 5.576917851028561e-05,
-      "loss": 1.6739,
-      "mean_token_accuracy": 0.6015461843907833,
+      "grad_norm": 0.21717922389507294,
+      "learning_rate": 5.5843908801869616e-05,
+      "loss": 1.6735,
+      "mean_token_accuracy": 0.6017058632075787,
       "step": 65500
     },
     {
       "epoch": 0.8967574067406265,
-      "grad_norm": 0.5268669128417969,
-      "learning_rate": 5.2372347074648765e-05,
+      "grad_norm": 0.3270646631717682,
+      "learning_rate": 5.244707736623278e-05,
       "loss": 1.6648,
-      "mean_token_accuracy": 0.6022867387533188,
+      "mean_token_accuracy": 0.6018540422320366,
       "step": 66000
     },
     {
       "epoch": 0.9035510234583585,
-      "grad_norm": 0.4265219271183014,
-      "learning_rate": 4.898910296475448e-05,
-      "loss": 1.6598,
-      "mean_token_accuracy": 0.6034302037358285,
+      "grad_norm": 0.24881796538829803,
+      "learning_rate": 4.906383325633849e-05,
+      "loss": 1.6592,
+      "mean_token_accuracy": 0.6032144545912742,
       "step": 66500
     },
     {
       "epoch": 0.9103446401760905,
-      "grad_norm": 0.43360719084739685,
-      "learning_rate": 4.5592271529117644e-05,
-      "loss": 1.6673,
-      "mean_token_accuracy": 0.6040239284336567,
+      "grad_norm": 0.22031113505363464,
+      "learning_rate": 4.566700182070165e-05,
+      "loss": 1.6669,
+      "mean_token_accuracy": 0.6039917682409286,
       "step": 67000
     },
     {
       "epoch": 0.9171382568938226,
-      "grad_norm": 0.6076446771621704,
-      "learning_rate": 4.220902741922335e-05,
-      "loss": 1.6684,
-      "mean_token_accuracy": 0.60401923969388,
+      "grad_norm": 0.3206457793712616,
+      "learning_rate": 4.2283757710807355e-05,
+      "loss": 1.6675,
+      "mean_token_accuracy": 0.6042173722684383,
       "step": 67500
     },
     {
       "epoch": 0.9239318736115546,
-      "grad_norm": 0.5381833910942078,
-      "learning_rate": 3.881219598358651e-05,
-      "loss": 1.6626,
-      "mean_token_accuracy": 0.6033281030356884,
+      "grad_norm": 0.2547774314880371,
+      "learning_rate": 3.8886926275170525e-05,
+      "loss": 1.6638,
+      "mean_token_accuracy": 0.6031933107674122,
       "step": 68000
     },
     {
       "epoch": 0.9307254903292866,
-      "grad_norm": 0.5558338165283203,
-      "learning_rate": 3.542895187369223e-05,
-      "loss": 1.6664,
-      "mean_token_accuracy": 0.6059176291823387,
+      "grad_norm": 0.2681569755077362,
+      "learning_rate": 3.5503682165276234e-05,
+      "loss": 1.6666,
+      "mean_token_accuracy": 0.6054518219828605,
       "step": 68500
     },
     {
       "epoch": 0.9375191070470186,
-      "grad_norm": 0.5438753962516785,
-      "learning_rate": 3.2032120438055383e-05,
-      "loss": 1.6533,
-      "mean_token_accuracy": 0.6052996191084385,
+      "grad_norm": 0.33085888624191284,
+      "learning_rate": 3.210685072963939e-05,
+      "loss": 1.6541,
+      "mean_token_accuracy": 0.6050065532922745,
       "step": 69000
     },
     {
       "epoch": 0.9443127237647506,
-      "grad_norm": 0.5837370753288269,
-      "learning_rate": 2.8648876328161093e-05,
-      "loss": 1.6775,
-      "mean_token_accuracy": 0.6013322997689247,
+      "grad_norm": 0.27647724747657776,
+      "learning_rate": 2.87236066197451e-05,
+      "loss": 1.6768,
+      "mean_token_accuracy": 0.6011287405788899,
       "step": 69500
     },
     {
       "epoch": 0.9511063404824827,
-      "grad_norm": 0.4310416877269745,
-      "learning_rate": 2.5252044892524253e-05,
-      "loss": 1.6548,
-      "mean_token_accuracy": 0.6055889129638672,
+      "grad_norm": 0.20769037306308746,
+      "learning_rate": 2.5326775184108264e-05,
+      "loss": 1.6556,
+      "mean_token_accuracy": 0.6048545605838299,
       "step": 70000
     },
     {
       "epoch": 0.9578999572002147,
-      "grad_norm": 0.5146716237068176,
-      "learning_rate": 2.1868800782629962e-05,
-      "loss": 1.6663,
-      "mean_token_accuracy": 0.6025079537630081,
+      "grad_norm": 0.22598762810230255,
+      "learning_rate": 2.1943531074213973e-05,
+      "loss": 1.6673,
+      "mean_token_accuracy": 0.6021151539683341,
       "step": 70500
     },
     {
       "epoch": 0.9646935739179467,
-      "grad_norm": 0.4616748094558716,
-      "learning_rate": 1.8471969346993126e-05,
-      "loss": 1.6703,
-      "mean_token_accuracy": 0.603231570750475,
+      "grad_norm": 0.22617878019809723,
+      "learning_rate": 1.8546699638577136e-05,
+      "loss": 1.6717,
+      "mean_token_accuracy": 0.6028414511382579,
       "step": 71000
     },
     {
       "epoch": 0.9714871906356787,
-      "grad_norm": 0.4664286673069,
-      "learning_rate": 1.5088725237098835e-05,
-      "loss": 1.6628,
-      "mean_token_accuracy": 0.6033680983781815,
+      "grad_norm": 0.21303826570510864,
+      "learning_rate": 1.5163455528682846e-05,
+      "loss": 1.664,
+      "mean_token_accuracy": 0.6029263936281204,
       "step": 71500
     },
     {
       "epoch": 0.9782808073534107,
-      "grad_norm": 0.4937819242477417,
-      "learning_rate": 1.1691893801461997e-05,
-      "loss": 1.6668,
-      "mean_token_accuracy": 0.6043202468752861,
+      "grad_norm": 0.23100796341896057,
+      "learning_rate": 1.1766624093046006e-05,
+      "loss": 1.6683,
+      "mean_token_accuracy": 0.6043952860236168,
       "step": 72000
     },
     {
       "epoch": 0.9850744240711428,
-      "grad_norm": 0.45690104365348816,
-      "learning_rate": 8.308649691567706e-06,
-      "loss": 1.66,
-      "mean_token_accuracy": 0.6045957347154617,
+      "grad_norm": 0.27368274331092834,
+      "learning_rate": 8.383379983151717e-06,
+      "loss": 1.6613,
+      "mean_token_accuracy": 0.604221729516983,
       "step": 72500
     },
     {
       "epoch": 0.9918680407888748,
-      "grad_norm": 0.43818211555480957,
-      "learning_rate": 4.911818255930868e-06,
-      "loss": 1.6669,
-      "mean_token_accuracy": 0.6046621772050857,
+      "grad_norm": 0.3219315707683563,
+      "learning_rate": 4.986548547514879e-06,
+      "loss": 1.669,
+      "mean_token_accuracy": 0.6041563397347927,
       "step": 73000
     },
     {
       "epoch": 0.9986616575066067,
-      "grad_norm": 0.4940794110298157,
-      "learning_rate": 1.528574146036577e-06,
-      "loss": 1.6517,
-      "mean_token_accuracy": 0.606719139367342,
+      "grad_norm": 0.2820158302783966,
+      "learning_rate": 1.6033044376205877e-06,
+      "loss": 1.6537,
+      "mean_token_accuracy": 0.6058071631491184,
       "step": 73500
     },
     {
       "epoch": 0.9999932063832823,
-      "mean_token_accuracy": 0.6021248713439825,
+      "mean_token_accuracy": 0.6008745219026294,
       "step": 73598,
-      "total_flos": 9.078073041438638e+17,
-      "train_loss": 1.6789688296530434,
-      "train_runtime": 12675.9905,
-      "train_samples_per_second": 23.224,
-      "train_steps_per_second": 5.806
+      "total_flos": 9.005567628877496e+17,
+      "train_loss": 1.67220352899737,
+      "train_runtime": 13481.0993,
+      "train_samples_per_second": 21.837,
+      "train_steps_per_second": 5.459
     }
   ],
   "logging_steps": 500,
@@ -1212,7 +1212,7 @@
       "attributes": {}
     }
   },
-  "total_flos": 9.078073041438638e+17,
+  "total_flos": 9.005567628877496e+17,
   "train_batch_size": 1,
   "trial_name": null,
   "trial_params": null