Model save

README.md

@@ -2,13 +2,11 @@
 license: other
 library_name: peft
 tags:
-- alignment-handbook
-- generated_from_trainer
 - trl
 - sft
 - generated_from_trainer
 datasets:
-- youndukn/ROLE_PLAY_INSTRUCT
+- generator
 base_model: Gryphe/MythoMax-L2-13b
 model-index:
 - name: mythomax-13b-sft-lora
@@ -20,9 +18,9 @@ should probably proofread and complete it, then remove this comment. -->
 
 # mythomax-13b-sft-lora
 
-This model is a fine-tuned version of [Gryphe/MythoMax-L2-13b](https://huggingface.co/Gryphe/MythoMax-L2-13b) on the youndukn/ROLE_PLAY_INSTRUCT dataset.
+This model is a fine-tuned version of [Gryphe/MythoMax-L2-13b](https://huggingface.co/Gryphe/MythoMax-L2-13b) on the generator dataset.
 It achieves the following results on the evaluation set:
-- Loss: 0.8712
+- Loss: 1.5062
 
 ## Model description
 
@@ -46,8 +44,8 @@ The following hyperparameters were used during training:
 - eval_batch_size: 8
 - seed: 42
 - distributed_type: multi-GPU
-- gradient_accumulation_steps: 2
-- total_train_batch_size: 8
+- gradient_accumulation_steps: 128
+- total_train_batch_size: 512
 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
 - lr_scheduler_type: cosine
 - lr_scheduler_warmup_ratio: 0.1
@@ -57,7 +55,7 @@ The following hyperparameters were used during training:
 
 | Training Loss | Epoch | Step | Validation Loss |
 |:-------------:|:-----:|:----:|:---------------:|
-| 0.8799        | 1.0   | 711  | 0.8712          |
+| 1.5495        | 0.99  | 11   | 1.5062          |
 
 
 ### Framework versions

adapter_config.json

@@ -15,17 +15,17 @@
   "megatron_core": "megatron.core",
   "modules_to_save": null,
   "peft_type": "LORA",
-  "r": 16,
+  "r": 64,
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
+    "q_proj",
     "v_proj",
+    "gate_proj",
     "down_proj",
     "k_proj",
-    "o_proj",
     "up_proj",
-    "gate_proj",
-    "q_proj"
+    "o_proj"
   ],
   "task_type": "CAUSAL_LM"
 }

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:55f703e544e39aa021e4f0bcd5a356fe1010a8795fdceaa684eca0e4df86c45e
-size 125249184
+oid sha256:0cb365c99f8cf3346fdbb14cd2122fe2728458bacd0b2b5bfb40312743dd1e18
+size 500771216

all_results.json

@@ -1,13 +1,13 @@
 {
-    "epoch": 1.0,
-    "eval_loss": 0.8712352514266968,
-    "eval_runtime": 636.6369,
+    "epoch": 0.99,
+    "eval_loss": 1.506221055984497,
+    "eval_runtime": 639.7589,
     "eval_samples": 1769,
-    "eval_samples_per_second": 0.996,
-    "eval_steps_per_second": 0.126,
-    "train_loss": 0.9507392455421587,
-    "train_runtime": 19724.4028,
+    "eval_samples_per_second": 0.991,
+    "eval_steps_per_second": 0.125,
+    "train_loss": 1.6525719924406572,
+    "train_runtime": 19611.737,
     "train_samples": 15899,
-    "train_samples_per_second": 0.288,
-    "train_steps_per_second": 0.036
+    "train_samples_per_second": 0.29,
+    "train_steps_per_second": 0.001
 }

eval_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "eval_loss": 0.8712352514266968,
-    "eval_runtime": 636.6369,
+    "epoch": 0.99,
+    "eval_loss": 1.506221055984497,
+    "eval_runtime": 639.7589,
     "eval_samples": 1769,
-    "eval_samples_per_second": 0.996,
-    "eval_steps_per_second": 0.126
+    "eval_samples_per_second": 0.991,
+    "eval_steps_per_second": 0.125
 }

runs/Jan17_14-36-44_77716d823da8/events.out.tfevents.1705502500.77716d823da8.825.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7d5fe914be043838003e537aaf1af37ccd3d5e94f3cf6fac99f71afb19338b3b
+size 4504

runs/Jan17_14-56-12_77716d823da8/events.out.tfevents.1705503412.77716d823da8.1051.0

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9cfd73f9dfb29b624f646973ab3a09dfac36b70b6653a6ccbfc63ce5885e59ce
+size 5580

runs/Jan17_14-56-12_77716d823da8/events.out.tfevents.1705523663.77716d823da8.1051.1

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:12cdaff80c2a7994e318ab692337a7bb10f5ba709bdea165889ee2efacea60b9
+size 354

train_results.json

@@ -1,8 +1,8 @@
 {
-    "epoch": 1.0,
-    "train_loss": 0.9507392455421587,
-    "train_runtime": 19724.4028,
+    "epoch": 0.99,
+    "train_loss": 1.6525719924406572,
+    "train_runtime": 19611.737,
     "train_samples": 15899,
-    "train_samples_per_second": 0.288,
-    "train_steps_per_second": 0.036
+    "train_samples_per_second": 0.29,
+    "train_steps_per_second": 0.001
 }

trainer_state.json

@@ -1,895 +1,55 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 0.9992972593113141,
+  "epoch": 0.9894588896697118,
   "eval_steps": 500,
-  "global_step": 711,
+  "global_step": 11,
   "is_hyper_param_search": false,
   "is_local_process_zero": true,
   "is_world_process_zero": true,
   "log_history": [
-    {
-      "epoch": 0.0,
-      "learning_rate": 2.777777777777778e-06,
-      "loss": 1.8297,
-      "step": 1
-    },
-    {
-      "epoch": 0.01,
-      "learning_rate": 1.388888888888889e-05,
-      "loss": 1.8176,
-      "step": 5
-    },
-    {
-      "epoch": 0.01,
-      "learning_rate": 2.777777777777778e-05,
-      "loss": 1.8447,
-      "step": 10
-    },
-    {
-      "epoch": 0.02,
-      "learning_rate": 4.166666666666667e-05,
-      "loss": 1.8067,
-      "step": 15
-    },
-    {
-      "epoch": 0.03,
-      "learning_rate": 5.555555555555556e-05,
-      "loss": 1.7216,
-      "step": 20
-    },
-    {
-      "epoch": 0.04,
-      "learning_rate": 6.944444444444444e-05,
-      "loss": 1.5794,
-      "step": 25
-    },
-    {
-      "epoch": 0.04,
-      "learning_rate": 8.333333333333334e-05,
-      "loss": 1.5115,
-      "step": 30
-    },
-    {
-      "epoch": 0.05,
-      "learning_rate": 9.722222222222223e-05,
-      "loss": 1.4649,
-      "step": 35
-    },
-    {
-      "epoch": 0.06,
-      "learning_rate": 0.00011111111111111112,
-      "loss": 1.4083,
-      "step": 40
-    },
-    {
-      "epoch": 0.06,
-      "learning_rate": 0.000125,
-      "loss": 1.3655,
-      "step": 45
-    },
-    {
-      "epoch": 0.07,
-      "learning_rate": 0.0001388888888888889,
-      "loss": 1.3277,
-      "step": 50
-    },
-    {
-      "epoch": 0.08,
-      "learning_rate": 0.00015277777777777777,
-      "loss": 1.3157,
-      "step": 55
-    },
-    {
-      "epoch": 0.08,
-      "learning_rate": 0.0001666666666666667,
-      "loss": 1.2691,
-      "step": 60
-    },
     {
       "epoch": 0.09,
-      "learning_rate": 0.00018055555555555557,
-      "loss": 1.25,
-      "step": 65
-    },
-    {
-      "epoch": 0.1,
-      "learning_rate": 0.00019444444444444446,
-      "loss": 1.2222,
-      "step": 70
-    },
-    {
-      "epoch": 0.11,
-      "learning_rate": 0.0001999891231617599,
-      "loss": 1.0998,
-      "step": 75
-    },
-    {
-      "epoch": 0.11,
-      "learning_rate": 0.00019992266216318035,
-      "loss": 0.9863,
-      "step": 80
-    },
-    {
-      "epoch": 0.12,
-      "learning_rate": 0.0001997958229642588,
-      "loss": 1.0027,
-      "step": 85
-    },
-    {
-      "epoch": 0.13,
-      "learning_rate": 0.00019960868220749448,
-      "loss": 0.9667,
-      "step": 90
-    },
-    {
-      "epoch": 0.13,
-      "learning_rate": 0.00019936135297256185,
-      "loss": 0.9511,
-      "step": 95
-    },
-    {
-      "epoch": 0.14,
-      "learning_rate": 0.00019905398470798207,
-      "loss": 0.9538,
-      "step": 100
-    },
-    {
-      "epoch": 0.15,
-      "learning_rate": 0.00019868676314081904,
-      "loss": 0.9471,
-      "step": 105
-    },
-    {
-      "epoch": 0.15,
-      "learning_rate": 0.00019825991016445387,
-      "loss": 0.9514,
-      "step": 110
-    },
-    {
-      "epoch": 0.16,
-      "learning_rate": 0.0001977736837045058,
-      "loss": 0.936,
-      "step": 115
-    },
-    {
-      "epoch": 0.17,
-      "learning_rate": 0.00019722837756298113,
-      "loss": 0.9268,
-      "step": 120
-    },
-    {
-      "epoch": 0.18,
-      "learning_rate": 0.00019662432124074327,
-      "loss": 0.93,
-      "step": 125
-    },
-    {
-      "epoch": 0.18,
-      "learning_rate": 0.00019596187973841217,
-      "loss": 0.9328,
-      "step": 130
-    },
-    {
-      "epoch": 0.19,
-      "learning_rate": 0.00019524145333581317,
-      "loss": 0.9228,
-      "step": 135
-    },
-    {
-      "epoch": 0.2,
-      "learning_rate": 0.00019446347735010763,
-      "loss": 0.9346,
-      "step": 140
-    },
-    {
-      "epoch": 0.2,
-      "learning_rate": 0.00019362842187275355,
-      "loss": 0.9255,
-      "step": 145
-    },
-    {
-      "epoch": 0.21,
-      "learning_rate": 0.00019273679148545245,
-      "loss": 0.9112,
-      "step": 150
-    },
-    {
-      "epoch": 0.22,
-      "learning_rate": 0.00019178912495525677,
-      "loss": 0.9124,
-      "step": 155
-    },
-    {
-      "epoch": 0.22,
-      "learning_rate": 0.00019078599490901984,
-      "loss": 0.9172,
-      "step": 160
-    },
-    {
-      "epoch": 0.23,
-      "learning_rate": 0.0001897280074873868,
-      "loss": 0.9099,
-      "step": 165
-    },
-    {
-      "epoch": 0.24,
-      "learning_rate": 0.00018861580197853422,
-      "loss": 0.9153,
-      "step": 170
-    },
-    {
-      "epoch": 0.25,
-      "learning_rate": 0.00018745005043188103,
-      "loss": 0.9142,
-      "step": 175
-    },
-    {
-      "epoch": 0.25,
-      "learning_rate": 0.00018623145725200278,
-      "loss": 0.9315,
-      "step": 180
-    },
-    {
-      "epoch": 0.26,
-      "learning_rate": 0.00018496075877299584,
-      "loss": 0.9145,
-      "step": 185
-    },
-    {
-      "epoch": 0.27,
-      "learning_rate": 0.00018363872281354797,
-      "loss": 0.8927,
-      "step": 190
-    },
-    {
-      "epoch": 0.27,
-      "learning_rate": 0.0001822661482129844,
-      "loss": 0.9047,
-      "step": 195
-    },
-    {
-      "epoch": 0.28,
-      "learning_rate": 0.0001808438643485698,
-      "loss": 0.8972,
-      "step": 200
-    },
-    {
-      "epoch": 0.29,
-      "learning_rate": 0.00017937273063435737,
-      "loss": 0.9032,
-      "step": 205
-    },
-    {
-      "epoch": 0.3,
-      "learning_rate": 0.00017785363600188894,
-      "loss": 0.9133,
-      "step": 210
-    },
-    {
-      "epoch": 0.3,
-      "learning_rate": 0.0001762874983630582,
-      "loss": 0.9044,
-      "step": 215
-    },
-    {
-      "epoch": 0.31,
-      "learning_rate": 0.00017467526405546343,
-      "loss": 0.8999,
-      "step": 220
-    },
-    {
-      "epoch": 0.32,
-      "learning_rate": 0.00017301790727058345,
-      "loss": 0.9027,
-      "step": 225
-    },
-    {
-      "epoch": 0.32,
-      "learning_rate": 0.00017131642946512313,
-      "loss": 0.8985,
-      "step": 230
-    },
-    {
-      "epoch": 0.33,
-      "learning_rate": 0.000169571858755884,
-      "loss": 0.909,
-      "step": 235
-    },
-    {
-      "epoch": 0.34,
-      "learning_rate": 0.00016778524929852512,
-      "loss": 0.8904,
-      "step": 240
-    },
-    {
-      "epoch": 0.34,
-      "learning_rate": 0.00016595768065059047,
-      "loss": 0.8983,
-      "step": 245
-    },
-    {
-      "epoch": 0.35,
-      "learning_rate": 0.0001640902571191869,
-      "loss": 0.8829,
-      "step": 250
-    },
-    {
-      "epoch": 0.36,
-      "learning_rate": 0.00016218410709370736,
-      "loss": 0.9014,
-      "step": 255
-    },
-    {
-      "epoch": 0.37,
-      "learning_rate": 0.00016024038236400246,
-      "loss": 0.8854,
-      "step": 260
-    },
-    {
-      "epoch": 0.37,
-      "learning_rate": 0.00015826025742441207,
-      "loss": 0.8947,
-      "step": 265
-    },
-    {
-      "epoch": 0.38,
-      "learning_rate": 0.0001562449287640781,
-      "loss": 0.894,
-      "step": 270
-    },
-    {
-      "epoch": 0.39,
-      "learning_rate": 0.00015419561414396657,
-      "loss": 0.898,
-      "step": 275
-    },
-    {
-      "epoch": 0.39,
-      "learning_rate": 0.00015211355186103655,
-      "loss": 0.9012,
-      "step": 280
-    },
-    {
-      "epoch": 0.4,
-      "learning_rate": 0.00015000000000000001,
-      "loss": 0.8836,
-      "step": 285
-    },
-    {
-      "epoch": 0.41,
-      "learning_rate": 0.00014785623567312492,
-      "loss": 0.8784,
-      "step": 290
-    },
-    {
-      "epoch": 0.41,
-      "learning_rate": 0.00014568355424854113,
-      "loss": 0.8905,
-      "step": 295
-    },
-    {
-      "epoch": 0.42,
-      "learning_rate": 0.00014348326856751496,
-      "loss": 0.8873,
-      "step": 300
-    },
-    {
-      "epoch": 0.43,
-      "learning_rate": 0.00014125670815116588,
-      "loss": 0.8773,
-      "step": 305
-    },
-    {
-      "epoch": 0.44,
-      "learning_rate": 0.00013900521839710426,
-      "loss": 0.9042,
-      "step": 310
-    },
-    {
-      "epoch": 0.44,
-      "learning_rate": 0.00013673015976647568,
-      "loss": 0.9007,
-      "step": 315
+      "learning_rate": 0.0001,
+      "loss": 1.8236,
+      "step": 1
     },
     {
       "epoch": 0.45,
-      "learning_rate": 0.00013443290696190334,
-      "loss": 0.8909,
-      "step": 320
-    },
-    {
-      "epoch": 0.46,
-      "learning_rate": 0.00013211484809682483,
-      "loss": 0.887,
-      "step": 325
-    },
-    {
-      "epoch": 0.46,
-      "learning_rate": 0.00012977738385672557,
-      "loss": 0.8729,
-      "step": 330
-    },
-    {
-      "epoch": 0.47,
-      "learning_rate": 0.00012742192665277568,
-      "loss": 0.8833,
-      "step": 335
-    },
-    {
-      "epoch": 0.48,
-      "learning_rate": 0.00012504989976838132,
-      "loss": 0.8891,
-      "step": 340
-    },
-    {
-      "epoch": 0.48,
-      "learning_rate": 0.0001226627364991667,
-      "loss": 0.884,
-      "step": 345
-    },
-    {
-      "epoch": 0.49,
-      "learning_rate": 0.00012026187928690629,
-      "loss": 0.886,
-      "step": 350
-    },
-    {
-      "epoch": 0.5,
-      "learning_rate": 0.00011784877884793031,
-      "loss": 0.8876,
-      "step": 355
-    },
-    {
-      "epoch": 0.51,
-      "learning_rate": 0.00011542489329653024,
-      "loss": 0.8865,
-      "step": 360
-    },
-    {
-      "epoch": 0.51,
-      "learning_rate": 0.00011299168726389448,
-      "loss": 0.8861,
-      "step": 365
-    },
-    {
-      "epoch": 0.52,
-      "learning_rate": 0.00011055063101310581,
-      "loss": 0.874,
-      "step": 370
-    },
-    {
-      "epoch": 0.53,
-      "learning_rate": 0.00010810319955073601,
-      "loss": 0.8807,
-      "step": 375
-    },
-    {
-      "epoch": 0.53,
-      "learning_rate": 0.00010565087173557395,
-      "loss": 0.8913,
-      "step": 380
-    },
-    {
-      "epoch": 0.54,
-      "learning_rate": 0.00010319512938502654,
-      "loss": 0.8937,
-      "step": 385
-    },
-    {
-      "epoch": 0.55,
-      "learning_rate": 0.00010073745637973124,
-      "loss": 0.8738,
-      "step": 390
-    },
-    {
-      "epoch": 0.56,
-      "learning_rate": 9.827933776692235e-05,
-      "loss": 0.8868,
-      "step": 395
-    },
-    {
-      "epoch": 0.56,
-      "learning_rate": 9.582225886309217e-05,
-      "loss": 0.8901,
-      "step": 400
-    },
-    {
-      "epoch": 0.57,
-      "learning_rate": 9.336770435648964e-05,
-      "loss": 0.881,
-      "step": 405
-    },
-    {
-      "epoch": 0.58,
-      "learning_rate": 9.091715740999828e-05,
-      "loss": 0.8922,
-      "step": 410
-    },
-    {
-      "epoch": 0.58,
-      "learning_rate": 8.84720987649363e-05,
-      "loss": 0.8695,
-      "step": 415
-    },
-    {
-      "epoch": 0.59,
-      "learning_rate": 8.60340058463194e-05,
-      "loss": 0.8959,
-      "step": 420
-    },
-    {
-      "epoch": 0.6,
-      "learning_rate": 8.360435187012788e-05,
-      "loss": 0.8757,
-      "step": 425
-    },
-    {
-      "epoch": 0.6,
-      "learning_rate": 8.118460495311686e-05,
-      "loss": 0.8832,
-      "step": 430
-    },
-    {
-      "epoch": 0.61,
-      "learning_rate": 7.877622722570771e-05,
-      "loss": 0.8807,
-      "step": 435
-    },
-    {
-      "epoch": 0.62,
-      "learning_rate": 7.638067394849671e-05,
-      "loss": 0.8664,
-      "step": 440
-    },
-    {
-      "epoch": 0.63,
-      "learning_rate": 7.399939263291493e-05,
-      "loss": 0.8539,
-      "step": 445
-    },
-    {
-      "epoch": 0.63,
-      "learning_rate": 7.163382216657034e-05,
-      "loss": 0.8616,
-      "step": 450
-    },
-    {
-      "epoch": 0.64,
-      "learning_rate": 6.928539194380102e-05,
-      "loss": 0.8619,
-      "step": 455
-    },
-    {
-      "epoch": 0.65,
-      "learning_rate": 6.695552100196452e-05,
-      "loss": 0.8686,
-      "step": 460
-    },
-    {
-      "epoch": 0.65,
-      "learning_rate": 6.464561716398565e-05,
-      "loss": 0.8732,
-      "step": 465
-    },
-    {
-      "epoch": 0.66,
-      "learning_rate": 6.235707618768032e-05,
-      "loss": 0.8701,
-      "step": 470
-    },
-    {
-      "epoch": 0.67,
-      "learning_rate": 6.009128092236983e-05,
-      "loss": 0.8812,
-      "step": 475
-    },
-    {
-      "epoch": 0.67,
-      "learning_rate": 5.784960047329519e-05,
-      "loss": 0.8758,
-      "step": 480
-    },
-    {
-      "epoch": 0.68,
-      "learning_rate": 5.563338937433622e-05,
-      "loss": 0.877,
-      "step": 485
-    },
-    {
-      "epoch": 0.69,
-      "learning_rate": 5.344398676953526e-05,
-      "loss": 0.8754,
-      "step": 490
-    },
-    {
-      "epoch": 0.7,
-      "learning_rate": 5.1282715603920374e-05,
-      "loss": 0.8702,
-      "step": 495
-    },
-    {
-      "epoch": 0.7,
-      "learning_rate": 4.915088182411675e-05,
-      "loss": 0.8798,
-      "step": 500
-    },
-    {
-      "epoch": 0.71,
-      "learning_rate": 4.7049773589229306e-05,
-      "loss": 0.8813,
-      "step": 505
-    },
-    {
-      "epoch": 0.72,
-      "learning_rate": 4.498066049247344e-05,
-      "loss": 0.8857,
-      "step": 510
-    },
-    {
-      "epoch": 0.72,
-      "learning_rate": 4.29447927940242e-05,
-      "loss": 0.8769,
-      "step": 515
-    },
-    {
-      "epoch": 0.73,
-      "learning_rate": 4.094340066554743e-05,
-      "loss": 0.8794,
-      "step": 520
-    },
-    {
-      "epoch": 0.74,
-      "learning_rate": 3.897769344686929e-05,
-      "loss": 0.8766,
-      "step": 525
-    },
-    {
-      "epoch": 0.74,
-      "learning_rate": 3.7048858915233664e-05,
-      "loss": 0.8678,
-      "step": 530
-    },
-    {
-      "epoch": 0.75,
-      "learning_rate": 3.515806256758847e-05,
-      "loss": 0.8538,
-      "step": 535
-    },
-    {
-      "epoch": 0.76,
-      "learning_rate": 3.330644691633492e-05,
-      "loss": 0.8692,
-      "step": 540
-    },
-    {
-      "epoch": 0.77,
-      "learning_rate": 3.149513079896521e-05,
-      "loss": 0.8638,
-      "step": 545
-    },
-    {
-      "epoch": 0.77,
-      "learning_rate": 2.9725208702005734e-05,
-      "loss": 0.8612,
-      "step": 550
-    },
-    {
-      "epoch": 0.78,
-      "learning_rate": 2.799775009967428e-05,
-      "loss": 0.872,
-      "step": 555
-    },
-    {
-      "epoch": 0.79,
-      "learning_rate": 2.631379880765107e-05,
-      "loss": 0.8836,
-      "step": 560
-    },
-    {
-      "epoch": 0.79,
-      "learning_rate": 2.4674372352353782e-05,
-      "loss": 0.8773,
-      "step": 565
-    },
-    {
-      "epoch": 0.8,
-      "learning_rate": 2.3080461356097937e-05,
-      "loss": 0.8609,
-      "step": 570
-    },
-    {
-      "epoch": 0.81,
-      "learning_rate": 2.1533028938514012e-05,
-      "loss": 0.8571,
-      "step": 575
-    },
-    {
-      "epoch": 0.82,
-      "learning_rate": 2.0033010134583086e-05,
-      "loss": 0.8782,
-      "step": 580
-    },
-    {
-      "epoch": 0.82,
-      "learning_rate": 1.858131132964259e-05,
-      "loss": 0.878,
-      "step": 585
-    },
-    {
-      "epoch": 0.83,
-      "learning_rate": 1.7178809711703523e-05,
-      "loss": 0.8755,
-      "step": 590
-    },
-    {
-      "epoch": 0.84,
-      "learning_rate": 1.5826352741410334e-05,
-      "loss": 0.8753,
-      "step": 595
-    },
-    {
-      "epoch": 0.84,
-      "learning_rate": 1.452475763996326e-05,
-      "loss": 0.8657,
-      "step": 600
-    },
-    {
-      "epoch": 0.85,
-      "learning_rate": 1.3274810895313083e-05,
-      "loss": 0.8544,
-      "step": 605
-    },
-    {
-      "epoch": 0.86,
-      "learning_rate": 1.207726778692625e-05,
-      "loss": 0.8521,
-      "step": 610
-    },
-    {
-      "epoch": 0.86,
-      "learning_rate": 1.0932851929407827e-05,
-      "loss": 0.8825,
-      "step": 615
-    },
-    {
-      "epoch": 0.87,
-      "learning_rate": 9.842254835257791e-06,
-      "loss": 0.8695,
-      "step": 620
-    },
-    {
-      "epoch": 0.88,
-      "learning_rate": 8.80613549702518e-06,
-      "loss": 0.8845,
-      "step": 625
-    },
-    {
-      "epoch": 0.89,
-      "learning_rate": 7.825119989112173e-06,
-      "loss": 0.8821,
-      "step": 630
-    },
-    {
-      "epoch": 0.89,
-      "learning_rate": 6.899801089469204e-06,
-      "loss": 0.8722,
-      "step": 635
+      "learning_rate": 0.00015000000000000001,
+      "loss": 1.775,
+      "step": 5
     },
     {
       "epoch": 0.9,
       "learning_rate": 6.030737921409169e-06,
-      "loss": 0.8653,
-      "step": 640
-    },
-    {
-      "epoch": 0.91,
-      "learning_rate": 5.2184556157576e-06,
-      "loss": 0.8717,
-      "step": 645
-    },
-    {
-      "epoch": 0.91,
-      "learning_rate": 4.463444993542721e-06,
-      "loss": 0.8811,
-      "step": 650
-    },
-    {
-      "epoch": 0.92,
-      "learning_rate": 3.7661622694171394e-06,
-      "loss": 0.8849,
-      "step": 655
-    },
-    {
-      "epoch": 0.93,
-      "learning_rate": 3.127028775990515e-06,
-      "loss": 0.8598,
-      "step": 660
-    },
-    {
-      "epoch": 0.93,
-      "learning_rate": 2.546430709239578e-06,
-      "loss": 0.8475,
-      "step": 665
-    },
-    {
-      "epoch": 0.94,
-      "learning_rate": 2.02471889514948e-06,
-      "loss": 0.8658,
-      "step": 670
-    },
-    {
-      "epoch": 0.95,
-      "learning_rate": 1.562208577727442e-06,
-      "loss": 0.8674,
-      "step": 675
-    },
-    {
-      "epoch": 0.96,
-      "learning_rate": 1.1591792285167603e-06,
-      "loss": 0.8774,
-      "step": 680
-    },
-    {
-      "epoch": 0.96,
-      "learning_rate": 8.158743777263333e-07,
-      "loss": 0.8786,
-      "step": 685
-    },
-    {
-      "epoch": 0.97,
-      "learning_rate": 5.325014670776951e-07,
-      "loss": 0.8766,
-      "step": 690
-    },
-    {
-      "epoch": 0.98,
-      "learning_rate": 3.092317244584919e-07,
-      "loss": 0.874,
-      "step": 695
-    },
-    {
-      "epoch": 0.98,
-      "learning_rate": 1.4620006045816815e-07,
-      "loss": 0.8683,
-      "step": 700
+      "loss": 1.5495,
+      "step": 10
     },
     {
       "epoch": 0.99,
-      "learning_rate": 4.350498684829729e-08,
-      "loss": 0.8548,
-      "step": 705
+      "eval_loss": 1.506221055984497,
+      "eval_runtime": 640.0172,
+      "eval_samples_per_second": 0.991,
+      "eval_steps_per_second": 0.125,
+      "step": 11
     },
     {
-      "epoch": 1.0,
-      "learning_rate": 1.20855705696421e-09,
-      "loss": 0.8799,
-      "step": 710
-    },
-    {
-      "epoch": 1.0,
-      "eval_loss": 0.8712352514266968,
-      "eval_runtime": 637.1016,
-      "eval_samples_per_second": 0.995,
-      "eval_steps_per_second": 0.126,
-      "step": 711
-    },
-    {
-      "epoch": 1.0,
-      "step": 711,
-      "total_flos": 9.02655718774014e+17,
-      "train_loss": 0.9507392455421587,
-      "train_runtime": 19724.4028,
-      "train_samples_per_second": 0.288,
-      "train_steps_per_second": 0.036
+      "epoch": 0.99,
+      "step": 11,
+      "total_flos": 9.067629788193096e+17,
+      "train_loss": 1.6525719924406572,
+      "train_runtime": 19611.737,
+      "train_samples_per_second": 0.29,
+      "train_steps_per_second": 0.001
     }
   ],
   "logging_steps": 5,
-  "max_steps": 711,
+  "max_steps": 11,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 1,
   "save_steps": 100,
-  "total_flos": 9.02655718774014e+17,
+  "total_flos": 9.067629788193096e+17,
   "train_batch_size": 4,
   "trial_name": null,
   "trial_params": null

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:e73049ec1c34175942c3be0595b94d5c7201a0e1f8ef51232c21bea6cb57bea8
+oid sha256:9d2eb55aad41f2b966da5a317095cc531f303dfdfe517a3735c14a12b6a6b483
 size 4792