Upload folder using huggingface_hub

README.md

@@ -1,16 +1,32 @@
 ---
-datasets:
-- yahma/alpaca-cleaned
-language:
-- en
-license: apache-2.0
-tags:
-- pytorch
-- causal-lm
-- llama2
-- fine-tuning
-- alpaca
-
+library_name: peft
 ---
+## Training procedure
+
+
+The following `bitsandbytes` quantization config was used during training:
+- load_in_8bit: True
+- load_in_4bit: False
+- llm_int8_threshold: 6.0
+- llm_int8_skip_modules: None
+- llm_int8_enable_fp32_cpu_offload: False
+- llm_int8_has_fp16_weight: False
+- bnb_4bit_quant_type: fp4
+- bnb_4bit_use_double_quant: False
+- bnb_4bit_compute_dtype: float32
+
+The following `bitsandbytes` quantization config was used during training:
+- load_in_8bit: True
+- load_in_4bit: False
+- llm_int8_threshold: 6.0
+- llm_int8_skip_modules: None
+- llm_int8_enable_fp32_cpu_offload: False
+- llm_int8_has_fp16_weight: False
+- bnb_4bit_quant_type: fp4
+- bnb_4bit_use_double_quant: False
+- bnb_4bit_compute_dtype: float32
+### Framework versions
+
+- PEFT 0.5.0.dev0
 
-# Llama-2-7b fine-tuned on LoRA alpaca-cleaned
+- PEFT 0.5.0.dev0

checkpoint-300/README.md

@@ -5,15 +5,15 @@ library_name: peft
 
 
 The following `bitsandbytes` quantization config was used during training:
-- load_in_8bit: False
-- load_in_4bit: True
+- load_in_8bit: True
+- load_in_4bit: False
 - llm_int8_threshold: 6.0
 - llm_int8_skip_modules: None
 - llm_int8_enable_fp32_cpu_offload: False
 - llm_int8_has_fp16_weight: False
-- bnb_4bit_quant_type: nf4
-- bnb_4bit_use_double_quant: True
-- bnb_4bit_compute_dtype: float16
+- bnb_4bit_quant_type: fp4
+- bnb_4bit_use_double_quant: False
+- bnb_4bit_compute_dtype: float32
 ### Framework versions
 
 

checkpoint-300/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3e194624690ed5347b1ee38663cafe4b90219064eeee73e8e90240e169d1c7a5
-size 12849093
+oid sha256:e50506142760259bd82a9c78175bec890b61a0e3fe649b140931d4f2ecea6bde
+size 50492421

checkpoint-300/rng_state.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:426b6985726cfa57b10a2ceb7852c6f85a389e9facba42145b01426ab7c985a4
+oid sha256:0226636856a5596306c2032091fb45bbfde326842fddb9676284dc01a2cd9733
 size 14575

checkpoint-300/scheduler.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:9850941e7721c4485c4be6cdc71482f604e4a686e50a540c13dcd9ef580d226b
+oid sha256:c8b283af915b370e85d1ae3d1dc0043f555ce603e26ec8b90b7113a4b578935a
 size 627

checkpoint-300/trainer_state.json

@@ -9,1808 +9,1808 @@
   "log_history": [
     {
       "epoch": 0.0,
-      "learning_rate": 3.333333333333333e-07,
-      "loss": 1.2235,
+      "learning_rate": 1.1111111111111112e-05,
+      "loss": 1.1222,
       "step": 1
     },
     {
       "epoch": 0.0,
-      "learning_rate": 6.666666666666666e-07,
-      "loss": 1.1855,
+      "learning_rate": 2.2222222222222223e-05,
+      "loss": 1.1901,
       "step": 2
     },
     {
       "epoch": 0.0,
-      "learning_rate": 1e-06,
-      "loss": 1.0787,
+      "learning_rate": 3.3333333333333335e-05,
+      "loss": 1.2159,
       "step": 3
     },
     {
       "epoch": 0.0,
-      "learning_rate": 1.3333333333333332e-06,
-      "loss": 1.2554,
+      "learning_rate": 4.4444444444444447e-05,
+      "loss": 1.204,
       "step": 4
     },
     {
       "epoch": 0.0,
-      "learning_rate": 1.6666666666666669e-06,
-      "loss": 1.303,
+      "learning_rate": 5.555555555555556e-05,
+      "loss": 1.2885,
       "step": 5
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2e-06,
-      "loss": 1.3766,
+      "learning_rate": 6.666666666666667e-05,
+      "loss": 1.1401,
       "step": 6
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.3333333333333336e-06,
-      "loss": 1.3295,
+      "learning_rate": 7.777777777777778e-05,
+      "loss": 1.1606,
       "step": 7
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.6666666666666664e-06,
-      "loss": 1.4096,
+      "learning_rate": 8.888888888888889e-05,
+      "loss": 1.3112,
       "step": 8
     },
     {
       "epoch": 0.0,
-      "learning_rate": 3e-06,
-      "loss": 1.2782,
+      "learning_rate": 0.0001,
+      "loss": 1.2529,
       "step": 9
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9999125880491853e-06,
-      "loss": 1.3555,
+      "learning_rate": 9.999708626830618e-05,
+      "loss": 1.1747,
       "step": 10
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9996503623845394e-06,
-      "loss": 1.4765,
+      "learning_rate": 9.998834541281798e-05,
+      "loss": 1.1645,
       "step": 11
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9992133535682728e-06,
-      "loss": 1.4239,
+      "learning_rate": 9.997377845227576e-05,
+      "loss": 1.3678,
       "step": 12
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.998601612533441e-06,
-      "loss": 1.4622,
+      "learning_rate": 9.995338708444804e-05,
+      "loss": 1.3055,
       "step": 13
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9978152105780155e-06,
-      "loss": 1.5584,
+      "learning_rate": 9.992717368593385e-05,
+      "loss": 1.2907,
       "step": 14
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9968542393565676e-06,
-      "loss": 1.4483,
+      "learning_rate": 9.989514131188559e-05,
+      "loss": 1.2025,
       "step": 15
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9957188108695897e-06,
-      "loss": 1.6347,
+      "learning_rate": 9.985729369565299e-05,
+      "loss": 1.3342,
       "step": 16
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.99440905745044e-06,
-      "loss": 1.5348,
+      "learning_rate": 9.9813635248348e-05,
+      "loss": 1.3678,
       "step": 17
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.992925131749921e-06,
-      "loss": 1.6787,
+      "learning_rate": 9.97641710583307e-05,
+      "loss": 1.273,
       "step": 18
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9912672067184863e-06,
-      "loss": 1.6322,
+      "learning_rate": 9.970890689061622e-05,
+      "loss": 1.2473,
       "step": 19
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9894354755860847e-06,
-      "loss": 1.6904,
+      "learning_rate": 9.964784918620282e-05,
+      "loss": 1.4571,
       "step": 20
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9874301518396377e-06,
-      "loss": 1.5149,
+      "learning_rate": 9.958100506132127e-05,
+      "loss": 1.2284,
       "step": 21
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.98525146919816e-06,
-      "loss": 1.6881,
+      "learning_rate": 9.950838230660534e-05,
+      "loss": 1.3918,
       "step": 22
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.982899681585518e-06,
-      "loss": 1.8038,
+      "learning_rate": 9.942998938618394e-05,
+      "loss": 1.1981,
       "step": 23
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.980375063100836e-06,
-      "loss": 1.7428,
+      "learning_rate": 9.934583543669453e-05,
+      "loss": 1.2575,
       "step": 24
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9776779079865498e-06,
-      "loss": 1.8379,
+      "learning_rate": 9.925593026621833e-05,
+      "loss": 1.3153,
       "step": 25
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9748085305941124e-06,
-      "loss": 1.8679,
+      "learning_rate": 9.916028435313708e-05,
+      "loss": 1.2294,
       "step": 26
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9717672653473587e-06,
-      "loss": 1.8585,
+      "learning_rate": 9.905890884491195e-05,
+      "loss": 1.1945,
       "step": 27
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9685544667035257e-06,
-      "loss": 1.787,
+      "learning_rate": 9.895181555678418e-05,
+      "loss": 1.1637,
       "step": 28
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9651705091119422e-06,
-      "loss": 1.9908,
+      "learning_rate": 9.883901697039808e-05,
+      "loss": 1.1551,
       "step": 29
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9616157869703894e-06,
-      "loss": 2.0559,
+      "learning_rate": 9.872052623234632e-05,
+      "loss": 1.0256,
       "step": 30
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.957890714579128e-06,
-      "loss": 2.1067,
+      "learning_rate": 9.85963571526376e-05,
+      "loss": 1.0165,
       "step": 31
     },
     {
       "epoch": 0.0,
-      "learning_rate": 2.9539957260926184e-06,
-      "loss": 2.1172,
+      "learning_rate": 9.846652420308728e-05,
+      "loss": 1.1713,
       "step": 32
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.949931275468917e-06,
-      "loss": 2.06,
+      "learning_rate": 9.833104251563056e-05,
+      "loss": 1.0292,
       "step": 33
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.9456978364167667e-06,
-      "loss": 2.2804,
+      "learning_rate": 9.818992788055889e-05,
+      "loss": 1.034,
       "step": 34
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.9412959023403906e-06,
-      "loss": 2.3487,
+      "learning_rate": 9.80431967446797e-05,
+      "loss": 1.0906,
       "step": 35
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.9367259862819805e-06,
-      "loss": 2.3993,
+      "learning_rate": 9.789086620939936e-05,
+      "loss": 1.0223,
       "step": 36
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.931988620861908e-06,
-      "loss": 2.4545,
+      "learning_rate": 9.773295402873026e-05,
+      "loss": 1.0137,
       "step": 37
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.9270843582166428e-06,
-      "loss": 2.3358,
+      "learning_rate": 9.756947860722143e-05,
+      "loss": 1.0469,
       "step": 38
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.9220137699344057e-06,
-      "loss": 2.4548,
+      "learning_rate": 9.740045899781352e-05,
+      "loss": 0.9361,
       "step": 39
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.9167774469885483e-06,
-      "loss": 2.4675,
+      "learning_rate": 9.722591489961827e-05,
+      "loss": 0.9553,
       "step": 40
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.911375999668675e-06,
-      "loss": 2.468,
+      "learning_rate": 9.70458666556225e-05,
+      "loss": 0.864,
       "step": 41
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.905810057509516e-06,
-      "loss": 2.3514,
+      "learning_rate": 9.686033525031719e-05,
+      "loss": 0.9793,
       "step": 42
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.900080269217554e-06,
-      "loss": 2.5067,
+      "learning_rate": 9.66693423072518e-05,
+      "loss": 0.8839,
       "step": 43
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8941873025954193e-06,
-      "loss": 2.5034,
+      "learning_rate": 9.647291008651398e-05,
+      "loss": 0.8898,
       "step": 44
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8881318444640566e-06,
-      "loss": 2.5753,
+      "learning_rate": 9.627106148213522e-05,
+      "loss": 0.7912,
       "step": 45
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.881914600582677e-06,
-      "loss": 2.5559,
+      "learning_rate": 9.606382001942255e-05,
+      "loss": 0.7436,
       "step": 46
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8755362955665014e-06,
-      "loss": 2.7476,
+      "learning_rate": 9.585120985221671e-05,
+      "loss": 0.7915,
       "step": 47
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8689976728023105e-06,
-      "loss": 2.6816,
+      "learning_rate": 9.563325576007701e-05,
+      "loss": 0.8672,
       "step": 48
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8622994943617984e-06,
-      "loss": 2.7578,
+      "learning_rate": 9.540998314539328e-05,
+      "loss": 0.7161,
       "step": 49
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.855442540912758e-06,
-      "loss": 2.8156,
+      "learning_rate": 9.518141803042527e-05,
+      "loss": 0.697,
       "step": 50
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.848427611628093e-06,
-      "loss": 1.1719,
+      "learning_rate": 9.494758705426978e-05,
+      "loss": 0.8822,
       "step": 51
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8412555240926745e-06,
-      "loss": 1.189,
+      "learning_rate": 9.470851746975582e-05,
+      "loss": 0.7704,
       "step": 52
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8339271142080537e-06,
-      "loss": 1.0689,
+      "learning_rate": 9.446423714026846e-05,
+      "loss": 0.8268,
       "step": 53
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8264432360950353e-06,
-      "loss": 1.1861,
+      "learning_rate": 9.421477453650118e-05,
+      "loss": 0.8874,
       "step": 54
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8188047619941344e-06,
-      "loss": 1.2506,
+      "learning_rate": 9.396015873313781e-05,
+      "loss": 0.8378,
       "step": 55
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.8110125821639135e-06,
-      "loss": 1.2313,
+      "learning_rate": 9.37004194054638e-05,
+      "loss": 0.889,
       "step": 56
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.803067604777227e-06,
-      "loss": 1.1761,
+      "learning_rate": 9.343558682590756e-05,
+      "loss": 0.9847,
       "step": 57
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.7949707558153703e-06,
-      "loss": 1.2601,
+      "learning_rate": 9.316569186051234e-05,
+      "loss": 0.9055,
       "step": 58
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.7867229789601615e-06,
-      "loss": 1.364,
+      "learning_rate": 9.289076596533872e-05,
+      "loss": 0.9204,
       "step": 59
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.778325235483954e-06,
-      "loss": 1.3515,
+      "learning_rate": 9.261084118279847e-05,
+      "loss": 0.9052,
       "step": 60
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.7697785041376006e-06,
-      "loss": 1.3414,
+      "learning_rate": 9.232595013792002e-05,
+      "loss": 0.8531,
       "step": 61
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.7610837810363814e-06,
-      "loss": 1.2057,
+      "learning_rate": 9.203612603454604e-05,
+      "loss": 0.8028,
       "step": 62
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.752242079543907e-06,
-      "loss": 1.404,
+      "learning_rate": 9.174140265146356e-05,
+      "loss": 0.926,
       "step": 63
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.743254430154012e-06,
-      "loss": 1.3719,
+      "learning_rate": 9.144181433846707e-05,
+      "loss": 0.9165,
       "step": 64
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.734121880370652e-06,
-      "loss": 1.4479,
+      "learning_rate": 9.113739601235507e-05,
+      "loss": 1.0392,
       "step": 65
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.7248454945858163e-06,
-      "loss": 1.4983,
+      "learning_rate": 9.082818315286055e-05,
+      "loss": 0.8548,
       "step": 66
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.7154263539554765e-06,
-      "loss": 1.477,
+      "learning_rate": 9.051421179851588e-05,
+      "loss": 0.898,
       "step": 67
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.7058655562735753e-06,
-      "loss": 1.5738,
+      "learning_rate": 9.01955185424525e-05,
+      "loss": 0.8343,
       "step": 68
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.696164215844081e-06,
-      "loss": 1.635,
+      "learning_rate": 8.987214052813604e-05,
+      "loss": 0.9507,
       "step": 69
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.6863234633511186e-06,
-      "loss": 1.541,
+      "learning_rate": 8.954411544503729e-05,
+      "loss": 0.8919,
       "step": 70
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.6763444457271836e-06,
-      "loss": 1.6623,
+      "learning_rate": 8.921148152423946e-05,
+      "loss": 0.9146,
       "step": 71
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.666228326019474e-06,
-      "loss": 1.7319,
+      "learning_rate": 8.887427753398248e-05,
+      "loss": 0.8907,
       "step": 72
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.655976283254334e-06,
-      "loss": 1.6835,
+      "learning_rate": 8.853254277514446e-05,
+      "loss": 0.8352,
       "step": 73
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.6455895122998405e-06,
-      "loss": 1.7268,
+      "learning_rate": 8.818631707666135e-05,
+      "loss": 0.972,
       "step": 74
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.6350692237265428e-06,
-      "loss": 1.7661,
+      "learning_rate": 8.783564079088477e-05,
+      "loss": 0.9635,
       "step": 75
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.624416643666371e-06,
-      "loss": 1.6734,
+      "learning_rate": 8.748055478887904e-05,
+      "loss": 0.843,
       "step": 76
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.6136330136697304e-06,
-      "loss": 1.802,
+      "learning_rate": 8.712110045565768e-05,
+      "loss": 0.8543,
       "step": 77
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.602719590560801e-06,
-      "loss": 1.8868,
+      "learning_rate": 8.675731968536002e-05,
+      "loss": 0.8737,
       "step": 78
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.591677646291054e-06,
-      "loss": 1.785,
+      "learning_rate": 8.638925487636848e-05,
+      "loss": 0.8705,
       "step": 79
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.58050846779101e-06,
-      "loss": 2.0649,
+      "learning_rate": 8.6016948926367e-05,
+      "loss": 1.0117,
       "step": 80
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.569213356820244e-06,
-      "loss": 1.9862,
+      "learning_rate": 8.564044522734147e-05,
+      "loss": 0.8361,
       "step": 81
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.557793629815669e-06,
-      "loss": 2.062,
+      "learning_rate": 8.52597876605223e-05,
+      "loss": 0.8629,
       "step": 82
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.5462506177381045e-06,
-      "loss": 1.9782,
+      "learning_rate": 8.487502059127015e-05,
+      "loss": 0.7811,
       "step": 83
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.5345856659171565e-06,
-      "loss": 2.1214,
+      "learning_rate": 8.448618886390522e-05,
+      "loss": 0.9086,
       "step": 84
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.522800133894418e-06,
-      "loss": 2.3212,
+      "learning_rate": 8.40933377964806e-05,
+      "loss": 0.7538,
       "step": 85
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.510895395265016e-06,
-      "loss": 2.2857,
+      "learning_rate": 8.369651317550054e-05,
+      "loss": 0.7346,
       "step": 86
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.498872837517522e-06,
-      "loss": 2.1979,
+      "learning_rate": 8.329576125058406e-05,
+      "loss": 0.9076,
       "step": 87
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.486733861872236e-06,
-      "loss": 2.332,
+      "learning_rate": 8.289112872907454e-05,
+      "loss": 0.8393,
       "step": 88
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.4744798831178817e-06,
-      "loss": 2.3778,
+      "learning_rate": 8.248266277059607e-05,
+      "loss": 0.8094,
       "step": 89
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.4621123294467098e-06,
-      "loss": 2.3099,
+      "learning_rate": 8.2070410981557e-05,
+      "loss": 0.8885,
       "step": 90
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.449632642288045e-06,
-      "loss": 2.4029,
+      "learning_rate": 8.16544214096015e-05,
+      "loss": 0.7709,
       "step": 91
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.437042276140287e-06,
-      "loss": 2.5416,
+      "learning_rate": 8.123474253800957e-05,
+      "loss": 0.7681,
       "step": 92
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.424342698401391e-06,
-      "loss": 2.531,
+      "learning_rate": 8.081142328004637e-05,
+      "loss": 0.7999,
       "step": 93
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.4115353891978432e-06,
-      "loss": 2.4672,
+      "learning_rate": 8.038451297326145e-05,
+      "loss": 0.8105,
       "step": 94
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.398621841212154e-06,
-      "loss": 2.4891,
+      "learning_rate": 7.995406137373846e-05,
+      "loss": 0.7392,
       "step": 95
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.3856035595088842e-06,
-      "loss": 2.5826,
+      "learning_rate": 7.952011865029614e-05,
+      "loss": 0.7033,
       "step": 96
     },
     {
       "epoch": 0.01,
-      "learning_rate": 2.372482061359234e-06,
-      "loss": 2.5583,
+      "learning_rate": 7.908273537864113e-05,
+      "loss": 0.689,
       "step": 97
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.3592588760642046e-06,
-      "loss": 2.5609,
+      "learning_rate": 7.86419625354735e-05,
+      "loss": 0.7046,
       "step": 98
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.34593554477636e-06,
-      "loss": 2.6917,
+      "learning_rate": 7.819785149254532e-05,
+      "loss": 0.6515,
       "step": 99
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.332513620320205e-06,
-      "loss": 2.6746,
+      "learning_rate": 7.77504540106735e-05,
+      "loss": 0.6036,
       "step": 100
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.318994667011207e-06,
-      "loss": 1.0723,
+      "learning_rate": 7.729982223370691e-05,
+      "loss": 0.8356,
       "step": 101
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.305380260473476e-06,
-      "loss": 1.1946,
+      "learning_rate": 7.68460086824492e-05,
+      "loss": 0.7933,
       "step": 102
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.2916719874561227e-06,
-      "loss": 1.1031,
+      "learning_rate": 7.638906624853743e-05,
+      "loss": 0.932,
       "step": 103
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.277871445648332e-06,
-      "loss": 1.2431,
+      "learning_rate": 7.592904818827775e-05,
+      "loss": 0.8673,
       "step": 104
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.2639802434931445e-06,
-      "loss": 1.1853,
+      "learning_rate": 7.546600811643816e-05,
+      "loss": 0.8731,
       "step": 105
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.25e-06,
-      "loss": 1.3308,
+      "learning_rate": 7.500000000000001e-05,
+      "loss": 0.8142,
       "step": 106
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.2359323445560408e-06,
-      "loss": 1.2904,
+      "learning_rate": 7.453107815186803e-05,
+      "loss": 0.8912,
       "step": 107
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.221778916736208e-06,
-      "loss": 1.338,
+      "learning_rate": 7.405929722454026e-05,
+      "loss": 0.8659,
       "step": 108
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.2075413661121492e-06,
-      "loss": 1.1788,
+      "learning_rate": 7.358471220373832e-05,
+      "loss": 0.9129,
       "step": 109
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.1932213520599652e-06,
-      "loss": 1.2099,
+      "learning_rate": 7.310737840199885e-05,
+      "loss": 0.7937,
       "step": 110
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.1788205435668085e-06,
-      "loss": 1.3613,
+      "learning_rate": 7.262735145222696e-05,
+      "loss": 0.8494,
       "step": 111
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.1643406190363625e-06,
-      "loss": 1.3139,
+      "learning_rate": 7.214468730121208e-05,
+      "loss": 0.9226,
       "step": 112
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.1497832660932298e-06,
-      "loss": 1.3749,
+      "learning_rate": 7.165944220310767e-05,
+      "loss": 0.7861,
       "step": 113
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.135150181386236e-06,
-      "loss": 1.3537,
+      "learning_rate": 7.117167271287453e-05,
+      "loss": 0.8482,
       "step": 114
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.1204430703906874e-06,
-      "loss": 1.4139,
+      "learning_rate": 7.068143567968957e-05,
+      "loss": 0.9018,
       "step": 115
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.1056636472096025e-06,
-      "loss": 1.5552,
+      "learning_rate": 7.018878824032009e-05,
+      "loss": 0.8878,
       "step": 116
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.090813634373931e-06,
-      "loss": 1.5982,
+      "learning_rate": 6.969378781246436e-05,
+      "loss": 0.9144,
       "step": 117
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.0758947626417945e-06,
-      "loss": 1.5037,
+      "learning_rate": 6.919649208805981e-05,
+      "loss": 0.9245,
       "step": 118
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.060908770796769e-06,
-      "loss": 1.6662,
+      "learning_rate": 6.869695902655897e-05,
+      "loss": 0.843,
       "step": 119
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.0458574054452316e-06,
-      "loss": 1.5392,
+      "learning_rate": 6.819524684817438e-05,
+      "loss": 0.9954,
       "step": 120
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.0307424208127912e-06,
-      "loss": 1.6647,
+      "learning_rate": 6.769141402709305e-05,
+      "loss": 0.9713,
       "step": 121
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.0155655785398396e-06,
-      "loss": 1.6124,
+      "learning_rate": 6.718551928466132e-05,
+      "loss": 1.0037,
       "step": 122
     },
     {
       "epoch": 0.02,
-      "learning_rate": 2.000328647476231e-06,
-      "loss": 1.674,
+      "learning_rate": 6.667762158254104e-05,
+      "loss": 0.6903,
       "step": 123
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.985033403475123e-06,
-      "loss": 1.674,
+      "learning_rate": 6.616778011583743e-05,
+      "loss": 0.8828,
       "step": 124
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.969681629186004e-06,
-      "loss": 1.6071,
+      "learning_rate": 6.565605430620013e-05,
+      "loss": 0.8054,
       "step": 125
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.954275113846926e-06,
-      "loss": 1.7843,
+      "learning_rate": 6.514250379489753e-05,
+      "loss": 0.8111,
       "step": 126
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.9388156530759715e-06,
-      "loss": 1.8194,
+      "learning_rate": 6.462718843586571e-05,
+      "loss": 0.9756,
       "step": 127
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.9233050486619715e-06,
-      "loss": 1.7772,
+      "learning_rate": 6.411016828873239e-05,
+      "loss": 1.0558,
       "step": 128
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.9077451083545143e-06,
-      "loss": 1.8367,
+      "learning_rate": 6.359150361181715e-05,
+      "loss": 0.886,
       "step": 129
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.8921376456532485e-06,
-      "loss": 1.9344,
+      "learning_rate": 6.307125485510828e-05,
+      "loss": 0.8109,
       "step": 130
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.8764844795965232e-06,
-      "loss": 1.8485,
+      "learning_rate": 6.254948265321744e-05,
+      "loss": 0.8609,
       "step": 131
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.8607874345493807e-06,
-      "loss": 1.9676,
+      "learning_rate": 6.202624781831268e-05,
+      "loss": 0.8583,
       "step": 132
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.8450483399909265e-06,
-      "loss": 1.9576,
+      "learning_rate": 6.150161133303089e-05,
+      "loss": 0.92,
       "step": 133
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.8292690303011076e-06,
-      "loss": 2.1381,
+      "learning_rate": 6.0975634343370256e-05,
+      "loss": 0.9136,
       "step": 134
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.813451344546913e-06,
-      "loss": 2.2337,
+      "learning_rate": 6.044837815156377e-05,
+      "loss": 0.8487,
       "step": 135
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.7975971262680348e-06,
-      "loss": 2.2179,
+      "learning_rate": 5.99199042089345e-05,
+      "loss": 0.8143,
       "step": 136
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.7817082232620054e-06,
-      "loss": 2.3267,
+      "learning_rate": 5.939027410873351e-05,
+      "loss": 0.8766,
       "step": 137
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.7657864873688345e-06,
-      "loss": 2.2561,
+      "learning_rate": 5.885954957896115e-05,
+      "loss": 0.7539,
       "step": 138
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.7498337742551817e-06,
-      "loss": 2.3142,
+      "learning_rate": 5.832779247517273e-05,
+      "loss": 0.7721,
       "step": 139
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.7338519431980798e-06,
-      "loss": 2.428,
+      "learning_rate": 5.779506477326933e-05,
+      "loss": 0.8023,
       "step": 140
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.7178428568682356e-06,
-      "loss": 2.4711,
+      "learning_rate": 5.726142856227452e-05,
+      "loss": 0.7718,
       "step": 141
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.701808381112938e-06,
-      "loss": 2.4576,
+      "learning_rate": 5.672694603709794e-05,
+      "loss": 0.8443,
       "step": 142
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.6857503847385956e-06,
-      "loss": 2.4396,
+      "learning_rate": 5.619167949128652e-05,
+      "loss": 0.8167,
       "step": 143
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.6696707392929268e-06,
-      "loss": 2.4912,
+      "learning_rate": 5.565569130976422e-05,
+      "loss": 0.8438,
       "step": 144
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.653571318846834e-06,
-      "loss": 2.5406,
+      "learning_rate": 5.5119043961561136e-05,
+      "loss": 0.6757,
       "step": 145
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.6374539997759822e-06,
-      "loss": 2.5187,
+      "learning_rate": 5.458179999253275e-05,
+      "loss": 0.6055,
       "step": 146
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.6213206605421064e-06,
-      "loss": 2.4131,
+      "learning_rate": 5.4044022018070214e-05,
+      "loss": 0.667,
       "step": 147
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.605173181474081e-06,
-      "loss": 2.5484,
+      "learning_rate": 5.3505772715802704e-05,
+      "loss": 0.6251,
       "step": 148
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.5890134445487679e-06,
-      "loss": 2.4955,
+      "learning_rate": 5.296711481829226e-05,
+      "loss": 0.6498,
       "step": 149
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.5728433331716726e-06,
-      "loss": 2.6633,
+      "learning_rate": 5.242811110572242e-05,
+      "loss": 0.5881,
       "step": 150
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.5566647319574351e-06,
-      "loss": 1.0591,
+      "learning_rate": 5.188882439858117e-05,
+      "loss": 0.7092,
       "step": 151
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.5404795265101808e-06,
-      "loss": 1.2533,
+      "learning_rate": 5.134931755033936e-05,
+      "loss": 0.8843,
       "step": 152
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.5242896032037523e-06,
-      "loss": 1.1398,
+      "learning_rate": 5.080965344012508e-05,
+      "loss": 0.8751,
       "step": 153
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.5080968489618567e-06,
-      "loss": 1.1983,
+      "learning_rate": 5.0269894965395225e-05,
+      "loss": 0.8272,
       "step": 154
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.4919031510381438e-06,
-      "loss": 1.2219,
+      "learning_rate": 4.973010503460479e-05,
+      "loss": 0.8899,
       "step": 155
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.4757103967962477e-06,
-      "loss": 1.2204,
+      "learning_rate": 4.919034655987493e-05,
+      "loss": 0.9079,
       "step": 156
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.4595204734898199e-06,
-      "loss": 1.2667,
+      "learning_rate": 4.865068244966066e-05,
+      "loss": 0.8265,
       "step": 157
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.4433352680425654e-06,
-      "loss": 1.1853,
+      "learning_rate": 4.8111175601418844e-05,
+      "loss": 0.8614,
       "step": 158
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.4271566668283281e-06,
-      "loss": 1.2905,
+      "learning_rate": 4.7571888894277604e-05,
+      "loss": 0.811,
       "step": 159
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.410986555451232e-06,
-      "loss": 1.4073,
+      "learning_rate": 4.703288518170774e-05,
+      "loss": 0.8822,
       "step": 160
     },
     {
       "epoch": 0.02,
-      "learning_rate": 1.3948268185259188e-06,
-      "loss": 1.3725,
+      "learning_rate": 4.6494227284197294e-05,
+      "loss": 0.876,
       "step": 161
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.3786793394578939e-06,
-      "loss": 1.3658,
+      "learning_rate": 4.59559779819298e-05,
+      "loss": 1.0022,
       "step": 162
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.362546000224018e-06,
-      "loss": 1.3991,
+      "learning_rate": 4.541820000746727e-05,
+      "loss": 0.8351,
       "step": 163
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.3464286811531663e-06,
-      "loss": 1.3649,
+      "learning_rate": 4.4880956038438876e-05,
+      "loss": 0.9567,
       "step": 164
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.3303292607070737e-06,
-      "loss": 1.4877,
+      "learning_rate": 4.434430869023579e-05,
+      "loss": 0.9064,
       "step": 165
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.314249615261405e-06,
-      "loss": 1.4299,
+      "learning_rate": 4.38083205087135e-05,
+      "loss": 0.817,
       "step": 166
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.2981916188870622e-06,
-      "loss": 1.3683,
+      "learning_rate": 4.3273053962902076e-05,
+      "loss": 0.8443,
       "step": 167
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.282157143131765e-06,
-      "loss": 1.473,
+      "learning_rate": 4.27385714377255e-05,
+      "loss": 0.8457,
       "step": 168
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.2661480568019203e-06,
-      "loss": 1.5913,
+      "learning_rate": 4.220493522673067e-05,
+      "loss": 0.8772,
       "step": 169
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.2501662257448184e-06,
-      "loss": 1.5045,
+      "learning_rate": 4.1672207524827275e-05,
+      "loss": 1.0726,
       "step": 170
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.234213512631166e-06,
-      "loss": 1.5012,
+      "learning_rate": 4.114045042103887e-05,
+      "loss": 0.9497,
       "step": 171
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.218291776737995e-06,
-      "loss": 1.5784,
+      "learning_rate": 4.06097258912665e-05,
+      "loss": 0.8689,
       "step": 172
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.2024028737319653e-06,
-      "loss": 1.5583,
+      "learning_rate": 4.0080095791065505e-05,
+      "loss": 0.9396,
       "step": 173
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.1865486554530874e-06,
-      "loss": 1.7044,
+      "learning_rate": 3.955162184843625e-05,
+      "loss": 0.9253,
       "step": 174
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.170730969698893e-06,
-      "loss": 1.7103,
+      "learning_rate": 3.902436565662977e-05,
+      "loss": 0.8986,
       "step": 175
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.154951660009074e-06,
-      "loss": 1.9039,
+      "learning_rate": 3.849838866696913e-05,
+      "loss": 0.9844,
       "step": 176
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.13921256545062e-06,
-      "loss": 1.8563,
+      "learning_rate": 3.7973752181687335e-05,
+      "loss": 0.9125,
       "step": 177
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.1235155204034768e-06,
-      "loss": 1.6943,
+      "learning_rate": 3.745051734678256e-05,
+      "loss": 0.8938,
       "step": 178
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.1078623543467518e-06,
-      "loss": 1.8663,
+      "learning_rate": 3.692874514489173e-05,
+      "loss": 0.8496,
       "step": 179
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.0922548916454855e-06,
-      "loss": 1.8081,
+      "learning_rate": 3.640849638818286e-05,
+      "loss": 0.9429,
       "step": 180
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.0766949513380286e-06,
-      "loss": 1.7813,
+      "learning_rate": 3.588983171126762e-05,
+      "loss": 0.8858,
       "step": 181
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.061184346924029e-06,
-      "loss": 2.0464,
+      "learning_rate": 3.53728115641343e-05,
+      "loss": 0.7386,
       "step": 182
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.0457248861530742e-06,
-      "loss": 1.9243,
+      "learning_rate": 3.4857496205102474e-05,
+      "loss": 1.0531,
       "step": 183
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.0303183708139966e-06,
-      "loss": 1.9261,
+      "learning_rate": 3.434394569379988e-05,
+      "loss": 0.9059,
       "step": 184
     },
     {
       "epoch": 0.03,
-      "learning_rate": 1.0149665965248775e-06,
-      "loss": 2.3096,
+      "learning_rate": 3.3832219884162585e-05,
+      "loss": 0.8182,
       "step": 185
     },
     {
       "epoch": 0.03,
-      "learning_rate": 9.996713525237694e-07,
-      "loss": 2.2348,
+      "learning_rate": 3.332237841745898e-05,
+      "loss": 0.7084,
       "step": 186
     },
     {
       "epoch": 0.03,
-      "learning_rate": 9.8443442146016e-07,
-      "loss": 2.1942,
+      "learning_rate": 3.281448071533867e-05,
+      "loss": 0.7416,
       "step": 187
     },
     {
       "epoch": 0.03,
-      "learning_rate": 9.69257579187209e-07,
-      "loss": 2.2085,
+      "learning_rate": 3.2308585972906966e-05,
+      "loss": 0.6783,
       "step": 188
     },
     {
       "epoch": 0.03,
-      "learning_rate": 9.54142594554769e-07,
-      "loss": 2.2091,
+      "learning_rate": 3.180475315182563e-05,
+      "loss": 0.6974,
       "step": 189
     },
     {
       "epoch": 0.03,
-      "learning_rate": 9.39091229203231e-07,
-      "loss": 2.1674,
+      "learning_rate": 3.130304097344103e-05,
+      "loss": 0.6956,
       "step": 190
     },
     {
       "epoch": 0.03,
-      "learning_rate": 9.241052373582058e-07,
-      "loss": 2.2736,
+      "learning_rate": 3.080350791194019e-05,
+      "loss": 0.6688,
       "step": 191
     },
     {
       "epoch": 0.03,
-      "learning_rate": 9.091863656260696e-07,
-      "loss": 2.2478,
+      "learning_rate": 3.0306212187535653e-05,
+      "loss": 0.662,
       "step": 192
     },
     {
       "epoch": 0.03,
-      "learning_rate": 8.943363527903977e-07,
-      "loss": 2.3178,
+      "learning_rate": 2.9811211759679924e-05,
+      "loss": 0.6405,
       "step": 193
     },
     {
       "epoch": 0.03,
-      "learning_rate": 8.795569296093133e-07,
-      "loss": 2.398,
+      "learning_rate": 2.9318564320310444e-05,
+      "loss": 0.5587,
       "step": 194
     },
     {
       "epoch": 0.03,
-      "learning_rate": 8.648498186137653e-07,
-      "loss": 2.489,
+      "learning_rate": 2.882832728712551e-05,
+      "loss": 0.6154,
       "step": 195
     },
     {
       "epoch": 0.03,
-      "learning_rate": 8.502167339067705e-07,
-      "loss": 2.3547,
+      "learning_rate": 2.8340557796892354e-05,
+      "loss": 0.6829,
       "step": 196
     },
     {
       "epoch": 0.03,
-      "learning_rate": 8.356593809636371e-07,
-      "loss": 2.3467,
+      "learning_rate": 2.7855312698787904e-05,
+      "loss": 0.5108,
       "step": 197
     },
     {
       "epoch": 0.03,
-      "learning_rate": 8.211794564331918e-07,
-      "loss": 2.4009,
+      "learning_rate": 2.737264854777306e-05,
+      "loss": 0.4153,
       "step": 198
     },
     {
       "epoch": 0.03,
-      "learning_rate": 8.067786479400346e-07,
-      "loss": 2.44,
+      "learning_rate": 2.6892621598001156e-05,
+      "loss": 0.474,
       "step": 199
     },
     {
       "epoch": 0.03,
-      "learning_rate": 7.924586338878512e-07,
-      "loss": 2.5891,
+      "learning_rate": 2.6415287796261706e-05,
+      "loss": 0.4005,
       "step": 200
     },
     {
       "epoch": 0.03,
-      "learning_rate": 7.782210832637924e-07,
-      "loss": 1.2204,
+      "learning_rate": 2.5940702775459747e-05,
+      "loss": 0.7317,
       "step": 201
     },
     {
       "epoch": 0.03,
-      "learning_rate": 7.640676554439594e-07,
-      "loss": 1.1149,
+      "learning_rate": 2.5468921848131983e-05,
+      "loss": 0.8145,
       "step": 202
     },
     {
       "epoch": 0.03,
-      "learning_rate": 7.500000000000003e-07,
-      "loss": 1.1802,
+      "learning_rate": 2.500000000000001e-05,
+      "loss": 0.8618,
       "step": 203
     },
     {
       "epoch": 0.03,
-      "learning_rate": 7.360197565068561e-07,
-      "loss": 1.0454,
+      "learning_rate": 2.4533991883561868e-05,
+      "loss": 0.9186,
       "step": 204
     },
     {
       "epoch": 0.03,
-      "learning_rate": 7.22128554351668e-07,
-      "loss": 1.1475,
+      "learning_rate": 2.407095181172227e-05,
+      "loss": 0.8658,
       "step": 205
     },
     {
       "epoch": 0.03,
-      "learning_rate": 7.083280125438766e-07,
-      "loss": 1.1811,
+      "learning_rate": 2.3610933751462553e-05,
+      "loss": 0.8398,
       "step": 206
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.946197395265243e-07,
-      "loss": 1.2031,
+      "learning_rate": 2.315399131755081e-05,
+      "loss": 0.8834,
       "step": 207
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.810053329887929e-07,
-      "loss": 1.3016,
+      "learning_rate": 2.2700177766293096e-05,
+      "loss": 0.9564,
       "step": 208
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.674863796797954e-07,
-      "loss": 1.1845,
+      "learning_rate": 2.2249545989326514e-05,
+      "loss": 0.8387,
       "step": 209
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.540644552236401e-07,
-      "loss": 1.2235,
+      "learning_rate": 2.180214850745467e-05,
+      "loss": 0.857,
       "step": 210
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.407411239357954e-07,
-      "loss": 1.3364,
+      "learning_rate": 2.1358037464526515e-05,
+      "loss": 0.8656,
       "step": 211
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.275179386407663e-07,
-      "loss": 1.2483,
+      "learning_rate": 2.091726462135888e-05,
+      "loss": 0.9432,
       "step": 212
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.143964404911165e-07,
-      "loss": 1.3154,
+      "learning_rate": 2.0479881349703883e-05,
+      "loss": 0.8754,
       "step": 213
     },
     {
       "epoch": 0.03,
-      "learning_rate": 6.013781587878464e-07,
-      "loss": 1.2914,
+      "learning_rate": 2.0045938626261546e-05,
+      "loss": 0.9271,
       "step": 214
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.884646108021563e-07,
-      "loss": 1.324,
+      "learning_rate": 1.9615487026738543e-05,
+      "loss": 0.9571,
       "step": 215
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.756573015986089e-07,
-      "loss": 1.382,
+      "learning_rate": 1.9188576719953633e-05,
+      "loss": 0.8626,
       "step": 216
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.629577238597132e-07,
-      "loss": 1.2145,
+      "learning_rate": 1.8765257461990442e-05,
+      "loss": 0.7676,
       "step": 217
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.503673577119552e-07,
-      "loss": 1.4172,
+      "learning_rate": 1.834557859039851e-05,
+      "loss": 0.9014,
       "step": 218
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.378876705532904e-07,
-      "loss": 1.4902,
+      "learning_rate": 1.7929589018443016e-05,
+      "loss": 0.8107,
       "step": 219
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.255201168821183e-07,
-      "loss": 1.5351,
+      "learning_rate": 1.7517337229403946e-05,
+      "loss": 0.7515,
       "step": 220
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.132661381277644e-07,
-      "loss": 1.4211,
+      "learning_rate": 1.710887127092548e-05,
+      "loss": 0.9335,
       "step": 221
     },
     {
       "epoch": 0.03,
-      "learning_rate": 5.011271624824787e-07,
-      "loss": 1.5289,
+      "learning_rate": 1.6704238749415957e-05,
+      "loss": 0.8645,
       "step": 222
     },
     {
       "epoch": 0.03,
-      "learning_rate": 4.891046047349837e-07,
-      "loss": 1.5527,
+      "learning_rate": 1.6303486824499458e-05,
+      "loss": 0.7009,
       "step": 223
     },
     {
       "epoch": 0.03,
-      "learning_rate": 4.771998661055823e-07,
-      "loss": 1.5057,
+      "learning_rate": 1.5906662203519412e-05,
+      "loss": 0.8673,
       "step": 224
     },
     {
       "epoch": 0.03,
-      "learning_rate": 4.6541433408284356e-07,
-      "loss": 1.5804,
+      "learning_rate": 1.5513811136094787e-05,
+      "loss": 0.8259,
       "step": 225
     },
     {
       "epoch": 0.03,
-      "learning_rate": 4.5374938226189584e-07,
-      "loss": 1.6606,
+      "learning_rate": 1.5124979408729861e-05,
+      "loss": 0.8709,
       "step": 226
     },
     {
       "epoch": 0.04,
-      "learning_rate": 4.4220637018433163e-07,
-      "loss": 1.6492,
+      "learning_rate": 1.4740212339477721e-05,
+      "loss": 0.7189,
       "step": 227
     },
     {
       "epoch": 0.04,
-      "learning_rate": 4.3078664317975654e-07,
-      "loss": 1.6729,
+      "learning_rate": 1.4359554772658552e-05,
+      "loss": 0.946,
       "step": 228
     },
     {
       "epoch": 0.04,
-      "learning_rate": 4.1949153220898987e-07,
-      "loss": 1.8831,
+      "learning_rate": 1.3983051073632997e-05,
+      "loss": 0.8452,
       "step": 229
     },
     {
       "epoch": 0.04,
-      "learning_rate": 4.0832235370894604e-07,
-      "loss": 1.8898,
+      "learning_rate": 1.3610745123631535e-05,
+      "loss": 0.7563,
       "step": 230
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.972804094391998e-07,
-      "loss": 1.6795,
+      "learning_rate": 1.3242680314639993e-05,
+      "loss": 0.738,
       "step": 231
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.863669863302698e-07,
-      "loss": 1.9394,
+      "learning_rate": 1.2878899544342327e-05,
+      "loss": 0.9574,
       "step": 232
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.755833563336293e-07,
-      "loss": 1.9454,
+      "learning_rate": 1.2519445211120979e-05,
+      "loss": 0.7155,
       "step": 233
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.64930776273457e-07,
-      "loss": 2.014,
+      "learning_rate": 1.2164359209115234e-05,
+      "loss": 0.7498,
       "step": 234
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.544104877001596e-07,
-      "loss": 1.9824,
+      "learning_rate": 1.1813682923338653e-05,
+      "loss": 0.8432,
       "step": 235
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.440237167456663e-07,
-      "loss": 2.0745,
+      "learning_rate": 1.1467457224855544e-05,
+      "loss": 0.729,
       "step": 236
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.337716739805264e-07,
-      "loss": 2.0536,
+      "learning_rate": 1.1125722466017547e-05,
+      "loss": 0.6728,
       "step": 237
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.2365555427281634e-07,
-      "loss": 2.2728,
+      "learning_rate": 1.0788518475760545e-05,
+      "loss": 0.6898,
       "step": 238
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.1367653664888173e-07,
-      "loss": 2.1603,
+      "learning_rate": 1.0455884554962725e-05,
+      "loss": 0.6286,
       "step": 239
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.0383578415591913e-07,
-      "loss": 2.1522,
+      "learning_rate": 1.012785947186397e-05,
+      "loss": 0.7136,
       "step": 240
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.9413444372642496e-07,
-      "loss": 2.2954,
+      "learning_rate": 9.804481457547498e-06,
+      "loss": 0.6081,
       "step": 241
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.8457364604452376e-07,
-      "loss": 2.1479,
+      "learning_rate": 9.485788201484126e-06,
+      "loss": 0.746,
       "step": 242
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.751545054141834e-07,
-      "loss": 2.3303,
+      "learning_rate": 9.171816847139448e-06,
+      "loss": 0.6714,
       "step": 243
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.6587811962934823e-07,
-      "loss": 2.3258,
+      "learning_rate": 8.86260398764494e-06,
+      "loss": 0.6298,
       "step": 244
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.567455698459882e-07,
-      "loss": 2.3271,
+      "learning_rate": 8.558185661532941e-06,
+      "loss": 0.5962,
       "step": 245
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.4775792045609353e-07,
-      "loss": 2.3152,
+      "learning_rate": 8.25859734853645e-06,
+      "loss": 0.6198,
       "step": 246
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.389162189636188e-07,
-      "loss": 2.4583,
+      "learning_rate": 7.96387396545396e-06,
+      "loss": 0.5893,
       "step": 247
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.3022149586239972e-07,
-      "loss": 2.3737,
+      "learning_rate": 7.67404986207999e-06,
+      "loss": 0.6294,
       "step": 248
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.2167476451604624e-07,
-      "loss": 2.5374,
+      "learning_rate": 7.389158817201542e-06,
+      "loss": 0.515,
       "step": 249
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.1327702103983864e-07,
-      "loss": 2.5436,
+      "learning_rate": 7.109234034661289e-06,
+      "loss": 0.4909,
       "step": 250
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.0502924418463014e-07,
-      "loss": 1.1269,
+      "learning_rate": 6.8343081394876715e-06,
+      "loss": 1.0134,
       "step": 251
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.9693239522277327e-07,
-      "loss": 1.1588,
+      "learning_rate": 6.564413174092443e-06,
+      "loss": 1.0413,
       "step": 252
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.8898741783608642e-07,
-      "loss": 1.127,
+      "learning_rate": 6.299580594536214e-06,
+      "loss": 0.7905,
       "step": 253
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.811952380058657e-07,
-      "loss": 1.1566,
+      "learning_rate": 6.0398412668621895e-06,
+      "loss": 0.7666,
       "step": 254
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.7355676390496482e-07,
-      "loss": 1.2204,
+      "learning_rate": 5.785225463498828e-06,
+      "loss": 0.8807,
       "step": 255
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.660728857919464e-07,
-      "loss": 1.1514,
+      "learning_rate": 5.535762859731547e-06,
+      "loss": 0.8202,
       "step": 256
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.5874447590732537e-07,
-      "loss": 1.1718,
+      "learning_rate": 5.291482530244179e-06,
+      "loss": 0.9371,
       "step": 257
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.5157238837190719e-07,
-      "loss": 1.3254,
+      "learning_rate": 5.05241294573024e-06,
+      "loss": 0.8992,
       "step": 258
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.4455745908724226e-07,
-      "loss": 1.2906,
+      "learning_rate": 4.818581969574742e-06,
+      "loss": 0.8329,
       "step": 259
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.377005056382018e-07,
-      "loss": 1.1925,
+      "learning_rate": 4.590016854606727e-06,
+      "loss": 0.9316,
       "step": 260
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.3100232719768996e-07,
-      "loss": 1.2717,
+      "learning_rate": 4.366744239922998e-06,
+      "loss": 0.8608,
       "step": 261
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.2446370443349863e-07,
-      "loss": 1.2414,
+      "learning_rate": 4.148790147783288e-06,
+      "loss": 0.9218,
       "step": 262
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.180853994173236e-07,
-      "loss": 1.2571,
+      "learning_rate": 3.936179980577453e-06,
+      "loss": 0.952,
       "step": 263
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.1186815553594382e-07,
-      "loss": 1.1211,
+      "learning_rate": 3.728938517864794e-06,
+      "loss": 0.8879,
       "step": 264
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.058126974045811e-07,
-      "loss": 1.3869,
+      "learning_rate": 3.527089913486037e-06,
+      "loss": 0.8573,
       "step": 265
     },
     {
       "epoch": 0.04,
-      "learning_rate": 9.991973078244638e-08,
-      "loss": 1.4214,
+      "learning_rate": 3.3306576927482126e-06,
+      "loss": 0.8729,
       "step": 266
     },
     {
       "epoch": 0.04,
-      "learning_rate": 9.418994249048474e-08,
-      "loss": 1.4063,
+      "learning_rate": 3.1396647496828247e-06,
+      "loss": 0.9766,
       "step": 267
     },
     {
       "epoch": 0.04,
-      "learning_rate": 8.862400033132573e-08,
-      "loss": 1.4621,
+      "learning_rate": 2.9541333443775243e-06,
+      "loss": 0.8435,
       "step": 268
     },
     {
       "epoch": 0.04,
-      "learning_rate": 8.322255301145204e-08,
-      "loss": 1.5625,
+      "learning_rate": 2.774085100381735e-06,
+      "loss": 0.7851,
       "step": 269
     },
     {
       "epoch": 0.04,
-      "learning_rate": 7.798623006559436e-08,
-      "loss": 1.519,
+      "learning_rate": 2.5995410021864787e-06,
+      "loss": 0.9089,
       "step": 270
     },
     {
       "epoch": 0.04,
-      "learning_rate": 7.291564178335719e-08,
-      "loss": 1.5473,
+      "learning_rate": 2.430521392778573e-06,
+      "loss": 0.8561,
       "step": 271
     },
     {
       "epoch": 0.04,
-      "learning_rate": 6.801137913809214e-08,
-      "loss": 1.5158,
+      "learning_rate": 2.2670459712697377e-06,
+      "loss": 0.8957,
       "step": 272
     },
     {
       "epoch": 0.04,
-      "learning_rate": 6.327401371801944e-08,
-      "loss": 1.6969,
+      "learning_rate": 2.1091337906006482e-06,
+      "loss": 0.7897,
       "step": 273
     },
     {
       "epoch": 0.04,
-      "learning_rate": 5.870409765960966e-08,
-      "loss": 1.6519,
+      "learning_rate": 1.956803255320322e-06,
+      "loss": 0.897,
       "step": 274
     },
     {
       "epoch": 0.04,
-      "learning_rate": 5.430216358323309e-08,
-      "loss": 1.5961,
+      "learning_rate": 1.810072119441103e-06,
+      "loss": 0.8682,
       "step": 275
     },
     {
       "epoch": 0.04,
-      "learning_rate": 5.00687245310833e-08,
-      "loss": 1.6128,
+      "learning_rate": 1.6689574843694433e-06,
+      "loss": 0.9593,
       "step": 276
     },
     {
       "epoch": 0.04,
-      "learning_rate": 4.60042739073816e-08,
-      "loss": 1.813,
+      "learning_rate": 1.53347579691272e-06,
+      "loss": 0.7004,
       "step": 277
     },
     {
       "epoch": 0.04,
-      "learning_rate": 4.2109285420872055e-08,
-      "loss": 1.6032,
+      "learning_rate": 1.4036428473624019e-06,
+      "loss": 0.9315,
       "step": 278
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.838421302961098e-08,
-      "loss": 1.8927,
+      "learning_rate": 1.2794737676536994e-06,
+      "loss": 0.8717,
       "step": 279
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.4829490888057424e-08,
-      "loss": 1.9264,
+      "learning_rate": 1.1609830296019143e-06,
+      "loss": 0.7959,
       "step": 280
     },
     {
       "epoch": 0.04,
-      "learning_rate": 3.1445533296474484e-08,
-      "loss": 1.7726,
+      "learning_rate": 1.0481844432158161e-06,
+      "loss": 0.8804,
       "step": 281
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.8232734652641424e-08,
-      "loss": 1.9217,
+      "learning_rate": 9.410911550880475e-07,
+      "loss": 0.7299,
       "step": 282
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.5191469405887625e-08,
-      "loss": 1.9583,
+      "learning_rate": 8.397156468629208e-07,
+      "loss": 0.733,
       "step": 283
     },
     {
       "epoch": 0.04,
-      "learning_rate": 2.2322092013450313e-08,
-      "loss": 2.0714,
+      "learning_rate": 7.44069733781677e-07,
+      "loss": 0.7353,
       "step": 284
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.962493689916395e-08,
-      "loss": 2.0708,
+      "learning_rate": 6.54164563305465e-07,
+      "loss": 0.7066,
       "step": 285
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.7100318414482063e-08,
-      "loss": 2.0683,
+      "learning_rate": 5.700106138160688e-07,
+      "loss": 0.7054,
       "step": 286
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.4748530801840076e-08,
-      "loss": 2.1105,
+      "learning_rate": 4.916176933946693e-07,
+      "loss": 0.8131,
       "step": 287
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.2569848160362384e-08,
-      "loss": 2.2285,
+      "learning_rate": 4.189949386787462e-07,
+      "loss": 0.6315,
       "step": 288
     },
     {
       "epoch": 0.04,
-      "learning_rate": 1.0564524413915422e-08,
-      "loss": 2.288,
+      "learning_rate": 3.5215081379718074e-07,
+      "loss": 0.6279,
       "step": 289
     },
     {
       "epoch": 0.04,
-      "learning_rate": 8.732793281513663e-09,
-      "loss": 2.2546,
+      "learning_rate": 2.9109310938378877e-07,
+      "loss": 0.6736,
       "step": 290
     },
     {
       "epoch": 0.04,
-      "learning_rate": 7.074868250079081e-09,
-      "loss": 2.156,
+      "learning_rate": 2.3582894166930268e-07,
+      "loss": 0.6892,
       "step": 291
     },
     {
       "epoch": 0.05,
-      "learning_rate": 5.590942549560052e-09,
-      "loss": 2.2368,
+      "learning_rate": 1.8636475165200174e-07,
+      "loss": 0.6936,
       "step": 292
     },
     {
       "epoch": 0.05,
-      "learning_rate": 4.2811891304105345e-09,
-      "loss": 2.2568,
+      "learning_rate": 1.427063043470178e-07,
+      "loss": 0.6369,
       "step": 293
     },
     {
       "epoch": 0.05,
-      "learning_rate": 3.145760643432527e-09,
-      "loss": 2.2339,
+      "learning_rate": 1.0485868811441757e-07,
+      "loss": 0.6448,
       "step": 294
     },
     {
       "epoch": 0.05,
-      "learning_rate": 2.1847894219846343e-09,
-      "loss": 2.3651,
+      "learning_rate": 7.282631406615447e-08,
+      "loss": 0.6009,
       "step": 295
     },
     {
       "epoch": 0.05,
-      "learning_rate": 1.3983874665589035e-09,
-      "loss": 2.492,
+      "learning_rate": 4.661291555196345e-08,
+      "loss": 0.5067,
       "step": 296
     },
     {
       "epoch": 0.05,
-      "learning_rate": 7.866464317276001e-10,
-      "loss": 2.3863,
+      "learning_rate": 2.6221547724253337e-08,
+      "loss": 0.5489,
       "step": 297
     },
     {
       "epoch": 0.05,
-      "learning_rate": 3.496376154604186e-10,
-      "loss": 2.4482,
+      "learning_rate": 1.1654587182013953e-08,
+      "loss": 0.5175,
       "step": 298
     },
     {
       "epoch": 0.05,
-      "learning_rate": 8.741195081479747e-11,
-      "loss": 2.5423,
+      "learning_rate": 2.9137316938265825e-09,
+      "loss": 0.5188,
       "step": 299
     },
     {
       "epoch": 0.05,
       "learning_rate": 0.0,
-      "loss": 2.5864,
+      "loss": 0.4117,
       "step": 300
     }
   ],
   "max_steps": 300,
   "num_train_epochs": 1,
-  "total_flos": 1.174396008923136e+16,
+  "total_flos": 2.3308777993273344e+16,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-300/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4373b7353f7618ba071e8149528a0ae191779985a891ab40089b4f1cda12cac9
+oid sha256:26d05e06b74b598e07338ab9a422f3e15db4fef4b5c96eb25a8e5404684e051b
 size 4027

flyte_training_config.json

@@ -1 +1 @@
-{"base_model": "meta-llama/Llama-2-7b-hf", "data_path": "yahma/alpaca-cleaned", "instruction_key": "instruction", "input_key": "input", "output_key": "output", "output_dir": "./output", "device_map": "auto", "batch_size": 8, "micro_batch_size": 8, "num_epochs": 1, "max_steps": 300, "eval_steps": 200, "save_steps": 50, "learning_rate": 3e-06, "cutoff_len": 512, "val_set_size": 0, "lora_r": 8, "lora_alpha": 16, "lora_dropout": 0.05, "weight_decay": 0.02, "warmup_ratio": 0.03, "lr_scheduler_type": "cosine", "lora_target_modules": ["q_proj", "k_proj", "v_proj"], "train_on_inputs": true, "add_eos_token": true, "group_by_length": true, "resume_from_checkpoint": null, "wandb_project": "unionai-llm-fine-tuning", "wandb_run_name": "", "wandb_watch": "", "wandb_log_model": "", "debug_mode": false, "debug_train_data_size": 1024}
+{"base_model": "meta-llama/Llama-2-7b-hf", "data_path": "yahma/alpaca-cleaned", "instruction_key": "instruction", "input_key": "input", "output_key": "output", "output_dir": "./output", "device_map": "auto", "batch_size": 8, "micro_batch_size": 8, "num_epochs": 1, "max_steps": 300, "eval_steps": 200, "save_steps": 50, "learning_rate": 0.0001, "cutoff_len": 512, "val_set_size": 0, "lora_r": 8, "lora_alpha": 16, "lora_dropout": 0.05, "weight_decay": 0.02, "warmup_ratio": 0.03, "lr_scheduler_type": "cosine", "lora_target_modules": ["q_proj", "k_proj", "v_proj"], "train_on_inputs": true, "add_eos_token": true, "group_by_length": true, "resume_from_checkpoint": null, "wandb_project": "unionai-llm-fine-tuning", "wandb_run_name": "", "wandb_watch": "", "wandb_log_model": "", "debug_mode": false, "debug_train_data_size": 1024}

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4373b7353f7618ba071e8149528a0ae191779985a891ab40089b4f1cda12cac9
+oid sha256:26d05e06b74b598e07338ab9a422f3e15db4fef4b5c96eb25a8e5404684e051b
 size 4027