Commit
•
dea9756
1
Parent(s):
2671109
Upload folder using huggingface_hub
Browse files- README.md +29 -13
- adapter_config.json +1 -0
- checkpoint-300/README.md +20 -0
- checkpoint-300/adapter_config.json +22 -0
- checkpoint-300/adapter_model.bin +3 -0
- checkpoint-300/adapter_model/README.md +20 -0
- checkpoint-300/adapter_model/adapter_config.json +22 -0
- checkpoint-300/adapter_model/adapter_model.bin +3 -0
- checkpoint-300/optimizer.pt +3 -0
- checkpoint-300/rng_state.pth +3 -0
- checkpoint-300/scheduler.pt +3 -0
- checkpoint-300/trainer_state.json +1816 -0
- checkpoint-300/training_args.bin +3 -0
- flyte_training_config.json +1 -1
- training_args.bin +2 -2
README.md
CHANGED
|
@@ -1,16 +1,32 @@
|
|
| 1 |
---
|
| 2 |
-
|
| 3 |
-
- yahma/alpaca-cleaned
|
| 4 |
-
language:
|
| 5 |
-
- en
|
| 6 |
-
license: apache-2.0
|
| 7 |
-
tags:
|
| 8 |
-
- pytorch
|
| 9 |
-
- causal-lm
|
| 10 |
-
- llama2
|
| 11 |
-
- fine-tuning
|
| 12 |
-
- alpaca
|
| 13 |
-
|
| 14 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
|
| 16 |
-
|
|
|
|
| 1 |
---
|
| 2 |
+
library_name: peft
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- load_in_8bit: False
|
| 9 |
+
- load_in_4bit: True
|
| 10 |
+
- llm_int8_threshold: 6.0
|
| 11 |
+
- llm_int8_skip_modules: None
|
| 12 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 13 |
+
- llm_int8_has_fp16_weight: False
|
| 14 |
+
- bnb_4bit_quant_type: nf4
|
| 15 |
+
- bnb_4bit_use_double_quant: True
|
| 16 |
+
- bnb_4bit_compute_dtype: float16
|
| 17 |
+
|
| 18 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 19 |
+
- load_in_8bit: False
|
| 20 |
+
- load_in_4bit: True
|
| 21 |
+
- llm_int8_threshold: 6.0
|
| 22 |
+
- llm_int8_skip_modules: None
|
| 23 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 24 |
+
- llm_int8_has_fp16_weight: False
|
| 25 |
+
- bnb_4bit_quant_type: nf4
|
| 26 |
+
- bnb_4bit_use_double_quant: True
|
| 27 |
+
- bnb_4bit_compute_dtype: float16
|
| 28 |
+
### Framework versions
|
| 29 |
+
|
| 30 |
+
- PEFT 0.5.0.dev0
|
| 31 |
|
| 32 |
+
- PEFT 0.5.0.dev0
|
adapter_config.json
CHANGED
|
@@ -1,4 +1,5 @@
|
|
| 1 |
{
|
|
|
|
| 2 |
"base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
|
| 3 |
"bias": "none",
|
| 4 |
"fan_in_fan_out": false,
|
|
|
|
| 1 |
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
"base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
|
| 4 |
"bias": "none",
|
| 5 |
"fan_in_fan_out": false,
|
checkpoint-300/README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- load_in_8bit: False
|
| 9 |
+
- load_in_4bit: True
|
| 10 |
+
- llm_int8_threshold: 6.0
|
| 11 |
+
- llm_int8_skip_modules: None
|
| 12 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 13 |
+
- llm_int8_has_fp16_weight: False
|
| 14 |
+
- bnb_4bit_quant_type: nf4
|
| 15 |
+
- bnb_4bit_use_double_quant: True
|
| 16 |
+
- bnb_4bit_compute_dtype: float16
|
| 17 |
+
### Framework versions
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
- PEFT 0.5.0.dev0
|
checkpoint-300/adapter_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 8,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"q_proj",
|
| 18 |
+
"k_proj",
|
| 19 |
+
"v_proj"
|
| 20 |
+
],
|
| 21 |
+
"task_type": "CAUSAL_LM"
|
| 22 |
+
}
|
checkpoint-300/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5e1621f48d9ad8feb1d6d31050275f0aafd080c5c07153301fe2f48411f4406
|
| 3 |
+
size 443
|
checkpoint-300/adapter_model/README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: peft
|
| 3 |
+
---
|
| 4 |
+
## Training procedure
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
The following `bitsandbytes` quantization config was used during training:
|
| 8 |
+
- load_in_8bit: False
|
| 9 |
+
- load_in_4bit: True
|
| 10 |
+
- llm_int8_threshold: 6.0
|
| 11 |
+
- llm_int8_skip_modules: None
|
| 12 |
+
- llm_int8_enable_fp32_cpu_offload: False
|
| 13 |
+
- llm_int8_has_fp16_weight: False
|
| 14 |
+
- bnb_4bit_quant_type: nf4
|
| 15 |
+
- bnb_4bit_use_double_quant: True
|
| 16 |
+
- bnb_4bit_compute_dtype: float16
|
| 17 |
+
### Framework versions
|
| 18 |
+
|
| 19 |
+
|
| 20 |
+
- PEFT 0.5.0.dev0
|
checkpoint-300/adapter_model/adapter_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"auto_mapping": null,
|
| 3 |
+
"base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
|
| 4 |
+
"bias": "none",
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"init_lora_weights": true,
|
| 8 |
+
"layers_pattern": null,
|
| 9 |
+
"layers_to_transform": null,
|
| 10 |
+
"lora_alpha": 16,
|
| 11 |
+
"lora_dropout": 0.05,
|
| 12 |
+
"modules_to_save": null,
|
| 13 |
+
"peft_type": "LORA",
|
| 14 |
+
"r": 8,
|
| 15 |
+
"revision": null,
|
| 16 |
+
"target_modules": [
|
| 17 |
+
"q_proj",
|
| 18 |
+
"k_proj",
|
| 19 |
+
"v_proj"
|
| 20 |
+
],
|
| 21 |
+
"task_type": "CAUSAL_LM"
|
| 22 |
+
}
|
checkpoint-300/adapter_model/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5e1621f48d9ad8feb1d6d31050275f0aafd080c5c07153301fe2f48411f4406
|
| 3 |
+
size 443
|
checkpoint-300/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:3e194624690ed5347b1ee38663cafe4b90219064eeee73e8e90240e169d1c7a5
|
| 3 |
+
size 12849093
|
checkpoint-300/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:426b6985726cfa57b10a2ceb7852c6f85a389e9facba42145b01426ab7c985a4
|
| 3 |
+
size 14575
|
checkpoint-300/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:9850941e7721c4485c4be6cdc71482f604e4a686e50a540c13dcd9ef580d226b
|
| 3 |
+
size 627
|
checkpoint-300/trainer_state.json
ADDED
|
@@ -0,0 +1,1816 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.04636785162287481,
|
| 5 |
+
"global_step": 300,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.0,
|
| 12 |
+
"learning_rate": 3.333333333333333e-07,
|
| 13 |
+
"loss": 1.2235,
|
| 14 |
+
"step": 1
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.0,
|
| 18 |
+
"learning_rate": 6.666666666666666e-07,
|
| 19 |
+
"loss": 1.1855,
|
| 20 |
+
"step": 2
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.0,
|
| 24 |
+
"learning_rate": 1e-06,
|
| 25 |
+
"loss": 1.0787,
|
| 26 |
+
"step": 3
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.0,
|
| 30 |
+
"learning_rate": 1.3333333333333332e-06,
|
| 31 |
+
"loss": 1.2554,
|
| 32 |
+
"step": 4
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.0,
|
| 36 |
+
"learning_rate": 1.6666666666666669e-06,
|
| 37 |
+
"loss": 1.303,
|
| 38 |
+
"step": 5
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.0,
|
| 42 |
+
"learning_rate": 2e-06,
|
| 43 |
+
"loss": 1.3766,
|
| 44 |
+
"step": 6
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.0,
|
| 48 |
+
"learning_rate": 2.3333333333333336e-06,
|
| 49 |
+
"loss": 1.3295,
|
| 50 |
+
"step": 7
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.0,
|
| 54 |
+
"learning_rate": 2.6666666666666664e-06,
|
| 55 |
+
"loss": 1.4096,
|
| 56 |
+
"step": 8
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.0,
|
| 60 |
+
"learning_rate": 3e-06,
|
| 61 |
+
"loss": 1.2782,
|
| 62 |
+
"step": 9
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.0,
|
| 66 |
+
"learning_rate": 2.9999125880491853e-06,
|
| 67 |
+
"loss": 1.3555,
|
| 68 |
+
"step": 10
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.0,
|
| 72 |
+
"learning_rate": 2.9996503623845394e-06,
|
| 73 |
+
"loss": 1.4765,
|
| 74 |
+
"step": 11
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.0,
|
| 78 |
+
"learning_rate": 2.9992133535682728e-06,
|
| 79 |
+
"loss": 1.4239,
|
| 80 |
+
"step": 12
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.0,
|
| 84 |
+
"learning_rate": 2.998601612533441e-06,
|
| 85 |
+
"loss": 1.4622,
|
| 86 |
+
"step": 13
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.0,
|
| 90 |
+
"learning_rate": 2.9978152105780155e-06,
|
| 91 |
+
"loss": 1.5584,
|
| 92 |
+
"step": 14
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.0,
|
| 96 |
+
"learning_rate": 2.9968542393565676e-06,
|
| 97 |
+
"loss": 1.4483,
|
| 98 |
+
"step": 15
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.0,
|
| 102 |
+
"learning_rate": 2.9957188108695897e-06,
|
| 103 |
+
"loss": 1.6347,
|
| 104 |
+
"step": 16
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.0,
|
| 108 |
+
"learning_rate": 2.99440905745044e-06,
|
| 109 |
+
"loss": 1.5348,
|
| 110 |
+
"step": 17
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.0,
|
| 114 |
+
"learning_rate": 2.992925131749921e-06,
|
| 115 |
+
"loss": 1.6787,
|
| 116 |
+
"step": 18
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.0,
|
| 120 |
+
"learning_rate": 2.9912672067184863e-06,
|
| 121 |
+
"loss": 1.6322,
|
| 122 |
+
"step": 19
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.0,
|
| 126 |
+
"learning_rate": 2.9894354755860847e-06,
|
| 127 |
+
"loss": 1.6904,
|
| 128 |
+
"step": 20
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.0,
|
| 132 |
+
"learning_rate": 2.9874301518396377e-06,
|
| 133 |
+
"loss": 1.5149,
|
| 134 |
+
"step": 21
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.0,
|
| 138 |
+
"learning_rate": 2.98525146919816e-06,
|
| 139 |
+
"loss": 1.6881,
|
| 140 |
+
"step": 22
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.0,
|
| 144 |
+
"learning_rate": 2.982899681585518e-06,
|
| 145 |
+
"loss": 1.8038,
|
| 146 |
+
"step": 23
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.0,
|
| 150 |
+
"learning_rate": 2.980375063100836e-06,
|
| 151 |
+
"loss": 1.7428,
|
| 152 |
+
"step": 24
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.0,
|
| 156 |
+
"learning_rate": 2.9776779079865498e-06,
|
| 157 |
+
"loss": 1.8379,
|
| 158 |
+
"step": 25
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.0,
|
| 162 |
+
"learning_rate": 2.9748085305941124e-06,
|
| 163 |
+
"loss": 1.8679,
|
| 164 |
+
"step": 26
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.0,
|
| 168 |
+
"learning_rate": 2.9717672653473587e-06,
|
| 169 |
+
"loss": 1.8585,
|
| 170 |
+
"step": 27
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.0,
|
| 174 |
+
"learning_rate": 2.9685544667035257e-06,
|
| 175 |
+
"loss": 1.787,
|
| 176 |
+
"step": 28
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.0,
|
| 180 |
+
"learning_rate": 2.9651705091119422e-06,
|
| 181 |
+
"loss": 1.9908,
|
| 182 |
+
"step": 29
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.0,
|
| 186 |
+
"learning_rate": 2.9616157869703894e-06,
|
| 187 |
+
"loss": 2.0559,
|
| 188 |
+
"step": 30
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.0,
|
| 192 |
+
"learning_rate": 2.957890714579128e-06,
|
| 193 |
+
"loss": 2.1067,
|
| 194 |
+
"step": 31
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.0,
|
| 198 |
+
"learning_rate": 2.9539957260926184e-06,
|
| 199 |
+
"loss": 2.1172,
|
| 200 |
+
"step": 32
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.01,
|
| 204 |
+
"learning_rate": 2.949931275468917e-06,
|
| 205 |
+
"loss": 2.06,
|
| 206 |
+
"step": 33
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.01,
|
| 210 |
+
"learning_rate": 2.9456978364167667e-06,
|
| 211 |
+
"loss": 2.2804,
|
| 212 |
+
"step": 34
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.01,
|
| 216 |
+
"learning_rate": 2.9412959023403906e-06,
|
| 217 |
+
"loss": 2.3487,
|
| 218 |
+
"step": 35
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.01,
|
| 222 |
+
"learning_rate": 2.9367259862819805e-06,
|
| 223 |
+
"loss": 2.3993,
|
| 224 |
+
"step": 36
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.01,
|
| 228 |
+
"learning_rate": 2.931988620861908e-06,
|
| 229 |
+
"loss": 2.4545,
|
| 230 |
+
"step": 37
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.01,
|
| 234 |
+
"learning_rate": 2.9270843582166428e-06,
|
| 235 |
+
"loss": 2.3358,
|
| 236 |
+
"step": 38
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.01,
|
| 240 |
+
"learning_rate": 2.9220137699344057e-06,
|
| 241 |
+
"loss": 2.4548,
|
| 242 |
+
"step": 39
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.01,
|
| 246 |
+
"learning_rate": 2.9167774469885483e-06,
|
| 247 |
+
"loss": 2.4675,
|
| 248 |
+
"step": 40
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.01,
|
| 252 |
+
"learning_rate": 2.911375999668675e-06,
|
| 253 |
+
"loss": 2.468,
|
| 254 |
+
"step": 41
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.01,
|
| 258 |
+
"learning_rate": 2.905810057509516e-06,
|
| 259 |
+
"loss": 2.3514,
|
| 260 |
+
"step": 42
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.01,
|
| 264 |
+
"learning_rate": 2.900080269217554e-06,
|
| 265 |
+
"loss": 2.5067,
|
| 266 |
+
"step": 43
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.01,
|
| 270 |
+
"learning_rate": 2.8941873025954193e-06,
|
| 271 |
+
"loss": 2.5034,
|
| 272 |
+
"step": 44
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.01,
|
| 276 |
+
"learning_rate": 2.8881318444640566e-06,
|
| 277 |
+
"loss": 2.5753,
|
| 278 |
+
"step": 45
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.01,
|
| 282 |
+
"learning_rate": 2.881914600582677e-06,
|
| 283 |
+
"loss": 2.5559,
|
| 284 |
+
"step": 46
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.01,
|
| 288 |
+
"learning_rate": 2.8755362955665014e-06,
|
| 289 |
+
"loss": 2.7476,
|
| 290 |
+
"step": 47
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.01,
|
| 294 |
+
"learning_rate": 2.8689976728023105e-06,
|
| 295 |
+
"loss": 2.6816,
|
| 296 |
+
"step": 48
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.01,
|
| 300 |
+
"learning_rate": 2.8622994943617984e-06,
|
| 301 |
+
"loss": 2.7578,
|
| 302 |
+
"step": 49
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.01,
|
| 306 |
+
"learning_rate": 2.855442540912758e-06,
|
| 307 |
+
"loss": 2.8156,
|
| 308 |
+
"step": 50
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.01,
|
| 312 |
+
"learning_rate": 2.848427611628093e-06,
|
| 313 |
+
"loss": 1.1719,
|
| 314 |
+
"step": 51
|
| 315 |
+
},
|
| 316 |
+
{
|
| 317 |
+
"epoch": 0.01,
|
| 318 |
+
"learning_rate": 2.8412555240926745e-06,
|
| 319 |
+
"loss": 1.189,
|
| 320 |
+
"step": 52
|
| 321 |
+
},
|
| 322 |
+
{
|
| 323 |
+
"epoch": 0.01,
|
| 324 |
+
"learning_rate": 2.8339271142080537e-06,
|
| 325 |
+
"loss": 1.0689,
|
| 326 |
+
"step": 53
|
| 327 |
+
},
|
| 328 |
+
{
|
| 329 |
+
"epoch": 0.01,
|
| 330 |
+
"learning_rate": 2.8264432360950353e-06,
|
| 331 |
+
"loss": 1.1861,
|
| 332 |
+
"step": 54
|
| 333 |
+
},
|
| 334 |
+
{
|
| 335 |
+
"epoch": 0.01,
|
| 336 |
+
"learning_rate": 2.8188047619941344e-06,
|
| 337 |
+
"loss": 1.2506,
|
| 338 |
+
"step": 55
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.01,
|
| 342 |
+
"learning_rate": 2.8110125821639135e-06,
|
| 343 |
+
"loss": 1.2313,
|
| 344 |
+
"step": 56
|
| 345 |
+
},
|
| 346 |
+
{
|
| 347 |
+
"epoch": 0.01,
|
| 348 |
+
"learning_rate": 2.803067604777227e-06,
|
| 349 |
+
"loss": 1.1761,
|
| 350 |
+
"step": 57
|
| 351 |
+
},
|
| 352 |
+
{
|
| 353 |
+
"epoch": 0.01,
|
| 354 |
+
"learning_rate": 2.7949707558153703e-06,
|
| 355 |
+
"loss": 1.2601,
|
| 356 |
+
"step": 58
|
| 357 |
+
},
|
| 358 |
+
{
|
| 359 |
+
"epoch": 0.01,
|
| 360 |
+
"learning_rate": 2.7867229789601615e-06,
|
| 361 |
+
"loss": 1.364,
|
| 362 |
+
"step": 59
|
| 363 |
+
},
|
| 364 |
+
{
|
| 365 |
+
"epoch": 0.01,
|
| 366 |
+
"learning_rate": 2.778325235483954e-06,
|
| 367 |
+
"loss": 1.3515,
|
| 368 |
+
"step": 60
|
| 369 |
+
},
|
| 370 |
+
{
|
| 371 |
+
"epoch": 0.01,
|
| 372 |
+
"learning_rate": 2.7697785041376006e-06,
|
| 373 |
+
"loss": 1.3414,
|
| 374 |
+
"step": 61
|
| 375 |
+
},
|
| 376 |
+
{
|
| 377 |
+
"epoch": 0.01,
|
| 378 |
+
"learning_rate": 2.7610837810363814e-06,
|
| 379 |
+
"loss": 1.2057,
|
| 380 |
+
"step": 62
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.01,
|
| 384 |
+
"learning_rate": 2.752242079543907e-06,
|
| 385 |
+
"loss": 1.404,
|
| 386 |
+
"step": 63
|
| 387 |
+
},
|
| 388 |
+
{
|
| 389 |
+
"epoch": 0.01,
|
| 390 |
+
"learning_rate": 2.743254430154012e-06,
|
| 391 |
+
"loss": 1.3719,
|
| 392 |
+
"step": 64
|
| 393 |
+
},
|
| 394 |
+
{
|
| 395 |
+
"epoch": 0.01,
|
| 396 |
+
"learning_rate": 2.734121880370652e-06,
|
| 397 |
+
"loss": 1.4479,
|
| 398 |
+
"step": 65
|
| 399 |
+
},
|
| 400 |
+
{
|
| 401 |
+
"epoch": 0.01,
|
| 402 |
+
"learning_rate": 2.7248454945858163e-06,
|
| 403 |
+
"loss": 1.4983,
|
| 404 |
+
"step": 66
|
| 405 |
+
},
|
| 406 |
+
{
|
| 407 |
+
"epoch": 0.01,
|
| 408 |
+
"learning_rate": 2.7154263539554765e-06,
|
| 409 |
+
"loss": 1.477,
|
| 410 |
+
"step": 67
|
| 411 |
+
},
|
| 412 |
+
{
|
| 413 |
+
"epoch": 0.01,
|
| 414 |
+
"learning_rate": 2.7058655562735753e-06,
|
| 415 |
+
"loss": 1.5738,
|
| 416 |
+
"step": 68
|
| 417 |
+
},
|
| 418 |
+
{
|
| 419 |
+
"epoch": 0.01,
|
| 420 |
+
"learning_rate": 2.696164215844081e-06,
|
| 421 |
+
"loss": 1.635,
|
| 422 |
+
"step": 69
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.01,
|
| 426 |
+
"learning_rate": 2.6863234633511186e-06,
|
| 427 |
+
"loss": 1.541,
|
| 428 |
+
"step": 70
|
| 429 |
+
},
|
| 430 |
+
{
|
| 431 |
+
"epoch": 0.01,
|
| 432 |
+
"learning_rate": 2.6763444457271836e-06,
|
| 433 |
+
"loss": 1.6623,
|
| 434 |
+
"step": 71
|
| 435 |
+
},
|
| 436 |
+
{
|
| 437 |
+
"epoch": 0.01,
|
| 438 |
+
"learning_rate": 2.666228326019474e-06,
|
| 439 |
+
"loss": 1.7319,
|
| 440 |
+
"step": 72
|
| 441 |
+
},
|
| 442 |
+
{
|
| 443 |
+
"epoch": 0.01,
|
| 444 |
+
"learning_rate": 2.655976283254334e-06,
|
| 445 |
+
"loss": 1.6835,
|
| 446 |
+
"step": 73
|
| 447 |
+
},
|
| 448 |
+
{
|
| 449 |
+
"epoch": 0.01,
|
| 450 |
+
"learning_rate": 2.6455895122998405e-06,
|
| 451 |
+
"loss": 1.7268,
|
| 452 |
+
"step": 74
|
| 453 |
+
},
|
| 454 |
+
{
|
| 455 |
+
"epoch": 0.01,
|
| 456 |
+
"learning_rate": 2.6350692237265428e-06,
|
| 457 |
+
"loss": 1.7661,
|
| 458 |
+
"step": 75
|
| 459 |
+
},
|
| 460 |
+
{
|
| 461 |
+
"epoch": 0.01,
|
| 462 |
+
"learning_rate": 2.624416643666371e-06,
|
| 463 |
+
"loss": 1.6734,
|
| 464 |
+
"step": 76
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.01,
|
| 468 |
+
"learning_rate": 2.6136330136697304e-06,
|
| 469 |
+
"loss": 1.802,
|
| 470 |
+
"step": 77
|
| 471 |
+
},
|
| 472 |
+
{
|
| 473 |
+
"epoch": 0.01,
|
| 474 |
+
"learning_rate": 2.602719590560801e-06,
|
| 475 |
+
"loss": 1.8868,
|
| 476 |
+
"step": 78
|
| 477 |
+
},
|
| 478 |
+
{
|
| 479 |
+
"epoch": 0.01,
|
| 480 |
+
"learning_rate": 2.591677646291054e-06,
|
| 481 |
+
"loss": 1.785,
|
| 482 |
+
"step": 79
|
| 483 |
+
},
|
| 484 |
+
{
|
| 485 |
+
"epoch": 0.01,
|
| 486 |
+
"learning_rate": 2.58050846779101e-06,
|
| 487 |
+
"loss": 2.0649,
|
| 488 |
+
"step": 80
|
| 489 |
+
},
|
| 490 |
+
{
|
| 491 |
+
"epoch": 0.01,
|
| 492 |
+
"learning_rate": 2.569213356820244e-06,
|
| 493 |
+
"loss": 1.9862,
|
| 494 |
+
"step": 81
|
| 495 |
+
},
|
| 496 |
+
{
|
| 497 |
+
"epoch": 0.01,
|
| 498 |
+
"learning_rate": 2.557793629815669e-06,
|
| 499 |
+
"loss": 2.062,
|
| 500 |
+
"step": 82
|
| 501 |
+
},
|
| 502 |
+
{
|
| 503 |
+
"epoch": 0.01,
|
| 504 |
+
"learning_rate": 2.5462506177381045e-06,
|
| 505 |
+
"loss": 1.9782,
|
| 506 |
+
"step": 83
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.01,
|
| 510 |
+
"learning_rate": 2.5345856659171565e-06,
|
| 511 |
+
"loss": 2.1214,
|
| 512 |
+
"step": 84
|
| 513 |
+
},
|
| 514 |
+
{
|
| 515 |
+
"epoch": 0.01,
|
| 516 |
+
"learning_rate": 2.522800133894418e-06,
|
| 517 |
+
"loss": 2.3212,
|
| 518 |
+
"step": 85
|
| 519 |
+
},
|
| 520 |
+
{
|
| 521 |
+
"epoch": 0.01,
|
| 522 |
+
"learning_rate": 2.510895395265016e-06,
|
| 523 |
+
"loss": 2.2857,
|
| 524 |
+
"step": 86
|
| 525 |
+
},
|
| 526 |
+
{
|
| 527 |
+
"epoch": 0.01,
|
| 528 |
+
"learning_rate": 2.498872837517522e-06,
|
| 529 |
+
"loss": 2.1979,
|
| 530 |
+
"step": 87
|
| 531 |
+
},
|
| 532 |
+
{
|
| 533 |
+
"epoch": 0.01,
|
| 534 |
+
"learning_rate": 2.486733861872236e-06,
|
| 535 |
+
"loss": 2.332,
|
| 536 |
+
"step": 88
|
| 537 |
+
},
|
| 538 |
+
{
|
| 539 |
+
"epoch": 0.01,
|
| 540 |
+
"learning_rate": 2.4744798831178817e-06,
|
| 541 |
+
"loss": 2.3778,
|
| 542 |
+
"step": 89
|
| 543 |
+
},
|
| 544 |
+
{
|
| 545 |
+
"epoch": 0.01,
|
| 546 |
+
"learning_rate": 2.4621123294467098e-06,
|
| 547 |
+
"loss": 2.3099,
|
| 548 |
+
"step": 90
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.01,
|
| 552 |
+
"learning_rate": 2.449632642288045e-06,
|
| 553 |
+
"loss": 2.4029,
|
| 554 |
+
"step": 91
|
| 555 |
+
},
|
| 556 |
+
{
|
| 557 |
+
"epoch": 0.01,
|
| 558 |
+
"learning_rate": 2.437042276140287e-06,
|
| 559 |
+
"loss": 2.5416,
|
| 560 |
+
"step": 92
|
| 561 |
+
},
|
| 562 |
+
{
|
| 563 |
+
"epoch": 0.01,
|
| 564 |
+
"learning_rate": 2.424342698401391e-06,
|
| 565 |
+
"loss": 2.531,
|
| 566 |
+
"step": 93
|
| 567 |
+
},
|
| 568 |
+
{
|
| 569 |
+
"epoch": 0.01,
|
| 570 |
+
"learning_rate": 2.4115353891978432e-06,
|
| 571 |
+
"loss": 2.4672,
|
| 572 |
+
"step": 94
|
| 573 |
+
},
|
| 574 |
+
{
|
| 575 |
+
"epoch": 0.01,
|
| 576 |
+
"learning_rate": 2.398621841212154e-06,
|
| 577 |
+
"loss": 2.4891,
|
| 578 |
+
"step": 95
|
| 579 |
+
},
|
| 580 |
+
{
|
| 581 |
+
"epoch": 0.01,
|
| 582 |
+
"learning_rate": 2.3856035595088842e-06,
|
| 583 |
+
"loss": 2.5826,
|
| 584 |
+
"step": 96
|
| 585 |
+
},
|
| 586 |
+
{
|
| 587 |
+
"epoch": 0.01,
|
| 588 |
+
"learning_rate": 2.372482061359234e-06,
|
| 589 |
+
"loss": 2.5583,
|
| 590 |
+
"step": 97
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.02,
|
| 594 |
+
"learning_rate": 2.3592588760642046e-06,
|
| 595 |
+
"loss": 2.5609,
|
| 596 |
+
"step": 98
|
| 597 |
+
},
|
| 598 |
+
{
|
| 599 |
+
"epoch": 0.02,
|
| 600 |
+
"learning_rate": 2.34593554477636e-06,
|
| 601 |
+
"loss": 2.6917,
|
| 602 |
+
"step": 99
|
| 603 |
+
},
|
| 604 |
+
{
|
| 605 |
+
"epoch": 0.02,
|
| 606 |
+
"learning_rate": 2.332513620320205e-06,
|
| 607 |
+
"loss": 2.6746,
|
| 608 |
+
"step": 100
|
| 609 |
+
},
|
| 610 |
+
{
|
| 611 |
+
"epoch": 0.02,
|
| 612 |
+
"learning_rate": 2.318994667011207e-06,
|
| 613 |
+
"loss": 1.0723,
|
| 614 |
+
"step": 101
|
| 615 |
+
},
|
| 616 |
+
{
|
| 617 |
+
"epoch": 0.02,
|
| 618 |
+
"learning_rate": 2.305380260473476e-06,
|
| 619 |
+
"loss": 1.1946,
|
| 620 |
+
"step": 102
|
| 621 |
+
},
|
| 622 |
+
{
|
| 623 |
+
"epoch": 0.02,
|
| 624 |
+
"learning_rate": 2.2916719874561227e-06,
|
| 625 |
+
"loss": 1.1031,
|
| 626 |
+
"step": 103
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.02,
|
| 630 |
+
"learning_rate": 2.277871445648332e-06,
|
| 631 |
+
"loss": 1.2431,
|
| 632 |
+
"step": 104
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.02,
|
| 636 |
+
"learning_rate": 2.2639802434931445e-06,
|
| 637 |
+
"loss": 1.1853,
|
| 638 |
+
"step": 105
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.02,
|
| 642 |
+
"learning_rate": 2.25e-06,
|
| 643 |
+
"loss": 1.3308,
|
| 644 |
+
"step": 106
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.02,
|
| 648 |
+
"learning_rate": 2.2359323445560408e-06,
|
| 649 |
+
"loss": 1.2904,
|
| 650 |
+
"step": 107
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.02,
|
| 654 |
+
"learning_rate": 2.221778916736208e-06,
|
| 655 |
+
"loss": 1.338,
|
| 656 |
+
"step": 108
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.02,
|
| 660 |
+
"learning_rate": 2.2075413661121492e-06,
|
| 661 |
+
"loss": 1.1788,
|
| 662 |
+
"step": 109
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.02,
|
| 666 |
+
"learning_rate": 2.1932213520599652e-06,
|
| 667 |
+
"loss": 1.2099,
|
| 668 |
+
"step": 110
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.02,
|
| 672 |
+
"learning_rate": 2.1788205435668085e-06,
|
| 673 |
+
"loss": 1.3613,
|
| 674 |
+
"step": 111
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.02,
|
| 678 |
+
"learning_rate": 2.1643406190363625e-06,
|
| 679 |
+
"loss": 1.3139,
|
| 680 |
+
"step": 112
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.02,
|
| 684 |
+
"learning_rate": 2.1497832660932298e-06,
|
| 685 |
+
"loss": 1.3749,
|
| 686 |
+
"step": 113
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.02,
|
| 690 |
+
"learning_rate": 2.135150181386236e-06,
|
| 691 |
+
"loss": 1.3537,
|
| 692 |
+
"step": 114
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.02,
|
| 696 |
+
"learning_rate": 2.1204430703906874e-06,
|
| 697 |
+
"loss": 1.4139,
|
| 698 |
+
"step": 115
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.02,
|
| 702 |
+
"learning_rate": 2.1056636472096025e-06,
|
| 703 |
+
"loss": 1.5552,
|
| 704 |
+
"step": 116
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.02,
|
| 708 |
+
"learning_rate": 2.090813634373931e-06,
|
| 709 |
+
"loss": 1.5982,
|
| 710 |
+
"step": 117
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.02,
|
| 714 |
+
"learning_rate": 2.0758947626417945e-06,
|
| 715 |
+
"loss": 1.5037,
|
| 716 |
+
"step": 118
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.02,
|
| 720 |
+
"learning_rate": 2.060908770796769e-06,
|
| 721 |
+
"loss": 1.6662,
|
| 722 |
+
"step": 119
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.02,
|
| 726 |
+
"learning_rate": 2.0458574054452316e-06,
|
| 727 |
+
"loss": 1.5392,
|
| 728 |
+
"step": 120
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.02,
|
| 732 |
+
"learning_rate": 2.0307424208127912e-06,
|
| 733 |
+
"loss": 1.6647,
|
| 734 |
+
"step": 121
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.02,
|
| 738 |
+
"learning_rate": 2.0155655785398396e-06,
|
| 739 |
+
"loss": 1.6124,
|
| 740 |
+
"step": 122
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.02,
|
| 744 |
+
"learning_rate": 2.000328647476231e-06,
|
| 745 |
+
"loss": 1.674,
|
| 746 |
+
"step": 123
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.02,
|
| 750 |
+
"learning_rate": 1.985033403475123e-06,
|
| 751 |
+
"loss": 1.674,
|
| 752 |
+
"step": 124
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.02,
|
| 756 |
+
"learning_rate": 1.969681629186004e-06,
|
| 757 |
+
"loss": 1.6071,
|
| 758 |
+
"step": 125
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.02,
|
| 762 |
+
"learning_rate": 1.954275113846926e-06,
|
| 763 |
+
"loss": 1.7843,
|
| 764 |
+
"step": 126
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.02,
|
| 768 |
+
"learning_rate": 1.9388156530759715e-06,
|
| 769 |
+
"loss": 1.8194,
|
| 770 |
+
"step": 127
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.02,
|
| 774 |
+
"learning_rate": 1.9233050486619715e-06,
|
| 775 |
+
"loss": 1.7772,
|
| 776 |
+
"step": 128
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.02,
|
| 780 |
+
"learning_rate": 1.9077451083545143e-06,
|
| 781 |
+
"loss": 1.8367,
|
| 782 |
+
"step": 129
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.02,
|
| 786 |
+
"learning_rate": 1.8921376456532485e-06,
|
| 787 |
+
"loss": 1.9344,
|
| 788 |
+
"step": 130
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.02,
|
| 792 |
+
"learning_rate": 1.8764844795965232e-06,
|
| 793 |
+
"loss": 1.8485,
|
| 794 |
+
"step": 131
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.02,
|
| 798 |
+
"learning_rate": 1.8607874345493807e-06,
|
| 799 |
+
"loss": 1.9676,
|
| 800 |
+
"step": 132
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.02,
|
| 804 |
+
"learning_rate": 1.8450483399909265e-06,
|
| 805 |
+
"loss": 1.9576,
|
| 806 |
+
"step": 133
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.02,
|
| 810 |
+
"learning_rate": 1.8292690303011076e-06,
|
| 811 |
+
"loss": 2.1381,
|
| 812 |
+
"step": 134
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.02,
|
| 816 |
+
"learning_rate": 1.813451344546913e-06,
|
| 817 |
+
"loss": 2.2337,
|
| 818 |
+
"step": 135
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.02,
|
| 822 |
+
"learning_rate": 1.7975971262680348e-06,
|
| 823 |
+
"loss": 2.2179,
|
| 824 |
+
"step": 136
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.02,
|
| 828 |
+
"learning_rate": 1.7817082232620054e-06,
|
| 829 |
+
"loss": 2.3267,
|
| 830 |
+
"step": 137
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.02,
|
| 834 |
+
"learning_rate": 1.7657864873688345e-06,
|
| 835 |
+
"loss": 2.2561,
|
| 836 |
+
"step": 138
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.02,
|
| 840 |
+
"learning_rate": 1.7498337742551817e-06,
|
| 841 |
+
"loss": 2.3142,
|
| 842 |
+
"step": 139
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.02,
|
| 846 |
+
"learning_rate": 1.7338519431980798e-06,
|
| 847 |
+
"loss": 2.428,
|
| 848 |
+
"step": 140
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.02,
|
| 852 |
+
"learning_rate": 1.7178428568682356e-06,
|
| 853 |
+
"loss": 2.4711,
|
| 854 |
+
"step": 141
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.02,
|
| 858 |
+
"learning_rate": 1.701808381112938e-06,
|
| 859 |
+
"loss": 2.4576,
|
| 860 |
+
"step": 142
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.02,
|
| 864 |
+
"learning_rate": 1.6857503847385956e-06,
|
| 865 |
+
"loss": 2.4396,
|
| 866 |
+
"step": 143
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.02,
|
| 870 |
+
"learning_rate": 1.6696707392929268e-06,
|
| 871 |
+
"loss": 2.4912,
|
| 872 |
+
"step": 144
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.02,
|
| 876 |
+
"learning_rate": 1.653571318846834e-06,
|
| 877 |
+
"loss": 2.5406,
|
| 878 |
+
"step": 145
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.02,
|
| 882 |
+
"learning_rate": 1.6374539997759822e-06,
|
| 883 |
+
"loss": 2.5187,
|
| 884 |
+
"step": 146
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.02,
|
| 888 |
+
"learning_rate": 1.6213206605421064e-06,
|
| 889 |
+
"loss": 2.4131,
|
| 890 |
+
"step": 147
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.02,
|
| 894 |
+
"learning_rate": 1.605173181474081e-06,
|
| 895 |
+
"loss": 2.5484,
|
| 896 |
+
"step": 148
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 0.02,
|
| 900 |
+
"learning_rate": 1.5890134445487679e-06,
|
| 901 |
+
"loss": 2.4955,
|
| 902 |
+
"step": 149
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 0.02,
|
| 906 |
+
"learning_rate": 1.5728433331716726e-06,
|
| 907 |
+
"loss": 2.6633,
|
| 908 |
+
"step": 150
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 0.02,
|
| 912 |
+
"learning_rate": 1.5566647319574351e-06,
|
| 913 |
+
"loss": 1.0591,
|
| 914 |
+
"step": 151
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 0.02,
|
| 918 |
+
"learning_rate": 1.5404795265101808e-06,
|
| 919 |
+
"loss": 1.2533,
|
| 920 |
+
"step": 152
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 0.02,
|
| 924 |
+
"learning_rate": 1.5242896032037523e-06,
|
| 925 |
+
"loss": 1.1398,
|
| 926 |
+
"step": 153
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 0.02,
|
| 930 |
+
"learning_rate": 1.5080968489618567e-06,
|
| 931 |
+
"loss": 1.1983,
|
| 932 |
+
"step": 154
|
| 933 |
+
},
|
| 934 |
+
{
|
| 935 |
+
"epoch": 0.02,
|
| 936 |
+
"learning_rate": 1.4919031510381438e-06,
|
| 937 |
+
"loss": 1.2219,
|
| 938 |
+
"step": 155
|
| 939 |
+
},
|
| 940 |
+
{
|
| 941 |
+
"epoch": 0.02,
|
| 942 |
+
"learning_rate": 1.4757103967962477e-06,
|
| 943 |
+
"loss": 1.2204,
|
| 944 |
+
"step": 156
|
| 945 |
+
},
|
| 946 |
+
{
|
| 947 |
+
"epoch": 0.02,
|
| 948 |
+
"learning_rate": 1.4595204734898199e-06,
|
| 949 |
+
"loss": 1.2667,
|
| 950 |
+
"step": 157
|
| 951 |
+
},
|
| 952 |
+
{
|
| 953 |
+
"epoch": 0.02,
|
| 954 |
+
"learning_rate": 1.4433352680425654e-06,
|
| 955 |
+
"loss": 1.1853,
|
| 956 |
+
"step": 158
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 0.02,
|
| 960 |
+
"learning_rate": 1.4271566668283281e-06,
|
| 961 |
+
"loss": 1.2905,
|
| 962 |
+
"step": 159
|
| 963 |
+
},
|
| 964 |
+
{
|
| 965 |
+
"epoch": 0.02,
|
| 966 |
+
"learning_rate": 1.410986555451232e-06,
|
| 967 |
+
"loss": 1.4073,
|
| 968 |
+
"step": 160
|
| 969 |
+
},
|
| 970 |
+
{
|
| 971 |
+
"epoch": 0.02,
|
| 972 |
+
"learning_rate": 1.3948268185259188e-06,
|
| 973 |
+
"loss": 1.3725,
|
| 974 |
+
"step": 161
|
| 975 |
+
},
|
| 976 |
+
{
|
| 977 |
+
"epoch": 0.03,
|
| 978 |
+
"learning_rate": 1.3786793394578939e-06,
|
| 979 |
+
"loss": 1.3658,
|
| 980 |
+
"step": 162
|
| 981 |
+
},
|
| 982 |
+
{
|
| 983 |
+
"epoch": 0.03,
|
| 984 |
+
"learning_rate": 1.362546000224018e-06,
|
| 985 |
+
"loss": 1.3991,
|
| 986 |
+
"step": 163
|
| 987 |
+
},
|
| 988 |
+
{
|
| 989 |
+
"epoch": 0.03,
|
| 990 |
+
"learning_rate": 1.3464286811531663e-06,
|
| 991 |
+
"loss": 1.3649,
|
| 992 |
+
"step": 164
|
| 993 |
+
},
|
| 994 |
+
{
|
| 995 |
+
"epoch": 0.03,
|
| 996 |
+
"learning_rate": 1.3303292607070737e-06,
|
| 997 |
+
"loss": 1.4877,
|
| 998 |
+
"step": 165
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 0.03,
|
| 1002 |
+
"learning_rate": 1.314249615261405e-06,
|
| 1003 |
+
"loss": 1.4299,
|
| 1004 |
+
"step": 166
|
| 1005 |
+
},
|
| 1006 |
+
{
|
| 1007 |
+
"epoch": 0.03,
|
| 1008 |
+
"learning_rate": 1.2981916188870622e-06,
|
| 1009 |
+
"loss": 1.3683,
|
| 1010 |
+
"step": 167
|
| 1011 |
+
},
|
| 1012 |
+
{
|
| 1013 |
+
"epoch": 0.03,
|
| 1014 |
+
"learning_rate": 1.282157143131765e-06,
|
| 1015 |
+
"loss": 1.473,
|
| 1016 |
+
"step": 168
|
| 1017 |
+
},
|
| 1018 |
+
{
|
| 1019 |
+
"epoch": 0.03,
|
| 1020 |
+
"learning_rate": 1.2661480568019203e-06,
|
| 1021 |
+
"loss": 1.5913,
|
| 1022 |
+
"step": 169
|
| 1023 |
+
},
|
| 1024 |
+
{
|
| 1025 |
+
"epoch": 0.03,
|
| 1026 |
+
"learning_rate": 1.2501662257448184e-06,
|
| 1027 |
+
"loss": 1.5045,
|
| 1028 |
+
"step": 170
|
| 1029 |
+
},
|
| 1030 |
+
{
|
| 1031 |
+
"epoch": 0.03,
|
| 1032 |
+
"learning_rate": 1.234213512631166e-06,
|
| 1033 |
+
"loss": 1.5012,
|
| 1034 |
+
"step": 171
|
| 1035 |
+
},
|
| 1036 |
+
{
|
| 1037 |
+
"epoch": 0.03,
|
| 1038 |
+
"learning_rate": 1.218291776737995e-06,
|
| 1039 |
+
"loss": 1.5784,
|
| 1040 |
+
"step": 172
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 0.03,
|
| 1044 |
+
"learning_rate": 1.2024028737319653e-06,
|
| 1045 |
+
"loss": 1.5583,
|
| 1046 |
+
"step": 173
|
| 1047 |
+
},
|
| 1048 |
+
{
|
| 1049 |
+
"epoch": 0.03,
|
| 1050 |
+
"learning_rate": 1.1865486554530874e-06,
|
| 1051 |
+
"loss": 1.7044,
|
| 1052 |
+
"step": 174
|
| 1053 |
+
},
|
| 1054 |
+
{
|
| 1055 |
+
"epoch": 0.03,
|
| 1056 |
+
"learning_rate": 1.170730969698893e-06,
|
| 1057 |
+
"loss": 1.7103,
|
| 1058 |
+
"step": 175
|
| 1059 |
+
},
|
| 1060 |
+
{
|
| 1061 |
+
"epoch": 0.03,
|
| 1062 |
+
"learning_rate": 1.154951660009074e-06,
|
| 1063 |
+
"loss": 1.9039,
|
| 1064 |
+
"step": 176
|
| 1065 |
+
},
|
| 1066 |
+
{
|
| 1067 |
+
"epoch": 0.03,
|
| 1068 |
+
"learning_rate": 1.13921256545062e-06,
|
| 1069 |
+
"loss": 1.8563,
|
| 1070 |
+
"step": 177
|
| 1071 |
+
},
|
| 1072 |
+
{
|
| 1073 |
+
"epoch": 0.03,
|
| 1074 |
+
"learning_rate": 1.1235155204034768e-06,
|
| 1075 |
+
"loss": 1.6943,
|
| 1076 |
+
"step": 178
|
| 1077 |
+
},
|
| 1078 |
+
{
|
| 1079 |
+
"epoch": 0.03,
|
| 1080 |
+
"learning_rate": 1.1078623543467518e-06,
|
| 1081 |
+
"loss": 1.8663,
|
| 1082 |
+
"step": 179
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 0.03,
|
| 1086 |
+
"learning_rate": 1.0922548916454855e-06,
|
| 1087 |
+
"loss": 1.8081,
|
| 1088 |
+
"step": 180
|
| 1089 |
+
},
|
| 1090 |
+
{
|
| 1091 |
+
"epoch": 0.03,
|
| 1092 |
+
"learning_rate": 1.0766949513380286e-06,
|
| 1093 |
+
"loss": 1.7813,
|
| 1094 |
+
"step": 181
|
| 1095 |
+
},
|
| 1096 |
+
{
|
| 1097 |
+
"epoch": 0.03,
|
| 1098 |
+
"learning_rate": 1.061184346924029e-06,
|
| 1099 |
+
"loss": 2.0464,
|
| 1100 |
+
"step": 182
|
| 1101 |
+
},
|
| 1102 |
+
{
|
| 1103 |
+
"epoch": 0.03,
|
| 1104 |
+
"learning_rate": 1.0457248861530742e-06,
|
| 1105 |
+
"loss": 1.9243,
|
| 1106 |
+
"step": 183
|
| 1107 |
+
},
|
| 1108 |
+
{
|
| 1109 |
+
"epoch": 0.03,
|
| 1110 |
+
"learning_rate": 1.0303183708139966e-06,
|
| 1111 |
+
"loss": 1.9261,
|
| 1112 |
+
"step": 184
|
| 1113 |
+
},
|
| 1114 |
+
{
|
| 1115 |
+
"epoch": 0.03,
|
| 1116 |
+
"learning_rate": 1.0149665965248775e-06,
|
| 1117 |
+
"loss": 2.3096,
|
| 1118 |
+
"step": 185
|
| 1119 |
+
},
|
| 1120 |
+
{
|
| 1121 |
+
"epoch": 0.03,
|
| 1122 |
+
"learning_rate": 9.996713525237694e-07,
|
| 1123 |
+
"loss": 2.2348,
|
| 1124 |
+
"step": 186
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 0.03,
|
| 1128 |
+
"learning_rate": 9.8443442146016e-07,
|
| 1129 |
+
"loss": 2.1942,
|
| 1130 |
+
"step": 187
|
| 1131 |
+
},
|
| 1132 |
+
{
|
| 1133 |
+
"epoch": 0.03,
|
| 1134 |
+
"learning_rate": 9.69257579187209e-07,
|
| 1135 |
+
"loss": 2.2085,
|
| 1136 |
+
"step": 188
|
| 1137 |
+
},
|
| 1138 |
+
{
|
| 1139 |
+
"epoch": 0.03,
|
| 1140 |
+
"learning_rate": 9.54142594554769e-07,
|
| 1141 |
+
"loss": 2.2091,
|
| 1142 |
+
"step": 189
|
| 1143 |
+
},
|
| 1144 |
+
{
|
| 1145 |
+
"epoch": 0.03,
|
| 1146 |
+
"learning_rate": 9.39091229203231e-07,
|
| 1147 |
+
"loss": 2.1674,
|
| 1148 |
+
"step": 190
|
| 1149 |
+
},
|
| 1150 |
+
{
|
| 1151 |
+
"epoch": 0.03,
|
| 1152 |
+
"learning_rate": 9.241052373582058e-07,
|
| 1153 |
+
"loss": 2.2736,
|
| 1154 |
+
"step": 191
|
| 1155 |
+
},
|
| 1156 |
+
{
|
| 1157 |
+
"epoch": 0.03,
|
| 1158 |
+
"learning_rate": 9.091863656260696e-07,
|
| 1159 |
+
"loss": 2.2478,
|
| 1160 |
+
"step": 192
|
| 1161 |
+
},
|
| 1162 |
+
{
|
| 1163 |
+
"epoch": 0.03,
|
| 1164 |
+
"learning_rate": 8.943363527903977e-07,
|
| 1165 |
+
"loss": 2.3178,
|
| 1166 |
+
"step": 193
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 0.03,
|
| 1170 |
+
"learning_rate": 8.795569296093133e-07,
|
| 1171 |
+
"loss": 2.398,
|
| 1172 |
+
"step": 194
|
| 1173 |
+
},
|
| 1174 |
+
{
|
| 1175 |
+
"epoch": 0.03,
|
| 1176 |
+
"learning_rate": 8.648498186137653e-07,
|
| 1177 |
+
"loss": 2.489,
|
| 1178 |
+
"step": 195
|
| 1179 |
+
},
|
| 1180 |
+
{
|
| 1181 |
+
"epoch": 0.03,
|
| 1182 |
+
"learning_rate": 8.502167339067705e-07,
|
| 1183 |
+
"loss": 2.3547,
|
| 1184 |
+
"step": 196
|
| 1185 |
+
},
|
| 1186 |
+
{
|
| 1187 |
+
"epoch": 0.03,
|
| 1188 |
+
"learning_rate": 8.356593809636371e-07,
|
| 1189 |
+
"loss": 2.3467,
|
| 1190 |
+
"step": 197
|
| 1191 |
+
},
|
| 1192 |
+
{
|
| 1193 |
+
"epoch": 0.03,
|
| 1194 |
+
"learning_rate": 8.211794564331918e-07,
|
| 1195 |
+
"loss": 2.4009,
|
| 1196 |
+
"step": 198
|
| 1197 |
+
},
|
| 1198 |
+
{
|
| 1199 |
+
"epoch": 0.03,
|
| 1200 |
+
"learning_rate": 8.067786479400346e-07,
|
| 1201 |
+
"loss": 2.44,
|
| 1202 |
+
"step": 199
|
| 1203 |
+
},
|
| 1204 |
+
{
|
| 1205 |
+
"epoch": 0.03,
|
| 1206 |
+
"learning_rate": 7.924586338878512e-07,
|
| 1207 |
+
"loss": 2.5891,
|
| 1208 |
+
"step": 200
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 0.03,
|
| 1212 |
+
"learning_rate": 7.782210832637924e-07,
|
| 1213 |
+
"loss": 1.2204,
|
| 1214 |
+
"step": 201
|
| 1215 |
+
},
|
| 1216 |
+
{
|
| 1217 |
+
"epoch": 0.03,
|
| 1218 |
+
"learning_rate": 7.640676554439594e-07,
|
| 1219 |
+
"loss": 1.1149,
|
| 1220 |
+
"step": 202
|
| 1221 |
+
},
|
| 1222 |
+
{
|
| 1223 |
+
"epoch": 0.03,
|
| 1224 |
+
"learning_rate": 7.500000000000003e-07,
|
| 1225 |
+
"loss": 1.1802,
|
| 1226 |
+
"step": 203
|
| 1227 |
+
},
|
| 1228 |
+
{
|
| 1229 |
+
"epoch": 0.03,
|
| 1230 |
+
"learning_rate": 7.360197565068561e-07,
|
| 1231 |
+
"loss": 1.0454,
|
| 1232 |
+
"step": 204
|
| 1233 |
+
},
|
| 1234 |
+
{
|
| 1235 |
+
"epoch": 0.03,
|
| 1236 |
+
"learning_rate": 7.22128554351668e-07,
|
| 1237 |
+
"loss": 1.1475,
|
| 1238 |
+
"step": 205
|
| 1239 |
+
},
|
| 1240 |
+
{
|
| 1241 |
+
"epoch": 0.03,
|
| 1242 |
+
"learning_rate": 7.083280125438766e-07,
|
| 1243 |
+
"loss": 1.1811,
|
| 1244 |
+
"step": 206
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 0.03,
|
| 1248 |
+
"learning_rate": 6.946197395265243e-07,
|
| 1249 |
+
"loss": 1.2031,
|
| 1250 |
+
"step": 207
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 0.03,
|
| 1254 |
+
"learning_rate": 6.810053329887929e-07,
|
| 1255 |
+
"loss": 1.3016,
|
| 1256 |
+
"step": 208
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 0.03,
|
| 1260 |
+
"learning_rate": 6.674863796797954e-07,
|
| 1261 |
+
"loss": 1.1845,
|
| 1262 |
+
"step": 209
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 0.03,
|
| 1266 |
+
"learning_rate": 6.540644552236401e-07,
|
| 1267 |
+
"loss": 1.2235,
|
| 1268 |
+
"step": 210
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 0.03,
|
| 1272 |
+
"learning_rate": 6.407411239357954e-07,
|
| 1273 |
+
"loss": 1.3364,
|
| 1274 |
+
"step": 211
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 0.03,
|
| 1278 |
+
"learning_rate": 6.275179386407663e-07,
|
| 1279 |
+
"loss": 1.2483,
|
| 1280 |
+
"step": 212
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 0.03,
|
| 1284 |
+
"learning_rate": 6.143964404911165e-07,
|
| 1285 |
+
"loss": 1.3154,
|
| 1286 |
+
"step": 213
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 0.03,
|
| 1290 |
+
"learning_rate": 6.013781587878464e-07,
|
| 1291 |
+
"loss": 1.2914,
|
| 1292 |
+
"step": 214
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 0.03,
|
| 1296 |
+
"learning_rate": 5.884646108021563e-07,
|
| 1297 |
+
"loss": 1.324,
|
| 1298 |
+
"step": 215
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 0.03,
|
| 1302 |
+
"learning_rate": 5.756573015986089e-07,
|
| 1303 |
+
"loss": 1.382,
|
| 1304 |
+
"step": 216
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 0.03,
|
| 1308 |
+
"learning_rate": 5.629577238597132e-07,
|
| 1309 |
+
"loss": 1.2145,
|
| 1310 |
+
"step": 217
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 0.03,
|
| 1314 |
+
"learning_rate": 5.503673577119552e-07,
|
| 1315 |
+
"loss": 1.4172,
|
| 1316 |
+
"step": 218
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 0.03,
|
| 1320 |
+
"learning_rate": 5.378876705532904e-07,
|
| 1321 |
+
"loss": 1.4902,
|
| 1322 |
+
"step": 219
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 0.03,
|
| 1326 |
+
"learning_rate": 5.255201168821183e-07,
|
| 1327 |
+
"loss": 1.5351,
|
| 1328 |
+
"step": 220
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 0.03,
|
| 1332 |
+
"learning_rate": 5.132661381277644e-07,
|
| 1333 |
+
"loss": 1.4211,
|
| 1334 |
+
"step": 221
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 0.03,
|
| 1338 |
+
"learning_rate": 5.011271624824787e-07,
|
| 1339 |
+
"loss": 1.5289,
|
| 1340 |
+
"step": 222
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 0.03,
|
| 1344 |
+
"learning_rate": 4.891046047349837e-07,
|
| 1345 |
+
"loss": 1.5527,
|
| 1346 |
+
"step": 223
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 0.03,
|
| 1350 |
+
"learning_rate": 4.771998661055823e-07,
|
| 1351 |
+
"loss": 1.5057,
|
| 1352 |
+
"step": 224
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 0.03,
|
| 1356 |
+
"learning_rate": 4.6541433408284356e-07,
|
| 1357 |
+
"loss": 1.5804,
|
| 1358 |
+
"step": 225
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 0.03,
|
| 1362 |
+
"learning_rate": 4.5374938226189584e-07,
|
| 1363 |
+
"loss": 1.6606,
|
| 1364 |
+
"step": 226
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 0.04,
|
| 1368 |
+
"learning_rate": 4.4220637018433163e-07,
|
| 1369 |
+
"loss": 1.6492,
|
| 1370 |
+
"step": 227
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 0.04,
|
| 1374 |
+
"learning_rate": 4.3078664317975654e-07,
|
| 1375 |
+
"loss": 1.6729,
|
| 1376 |
+
"step": 228
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 0.04,
|
| 1380 |
+
"learning_rate": 4.1949153220898987e-07,
|
| 1381 |
+
"loss": 1.8831,
|
| 1382 |
+
"step": 229
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 0.04,
|
| 1386 |
+
"learning_rate": 4.0832235370894604e-07,
|
| 1387 |
+
"loss": 1.8898,
|
| 1388 |
+
"step": 230
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 0.04,
|
| 1392 |
+
"learning_rate": 3.972804094391998e-07,
|
| 1393 |
+
"loss": 1.6795,
|
| 1394 |
+
"step": 231
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 0.04,
|
| 1398 |
+
"learning_rate": 3.863669863302698e-07,
|
| 1399 |
+
"loss": 1.9394,
|
| 1400 |
+
"step": 232
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 0.04,
|
| 1404 |
+
"learning_rate": 3.755833563336293e-07,
|
| 1405 |
+
"loss": 1.9454,
|
| 1406 |
+
"step": 233
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 0.04,
|
| 1410 |
+
"learning_rate": 3.64930776273457e-07,
|
| 1411 |
+
"loss": 2.014,
|
| 1412 |
+
"step": 234
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 0.04,
|
| 1416 |
+
"learning_rate": 3.544104877001596e-07,
|
| 1417 |
+
"loss": 1.9824,
|
| 1418 |
+
"step": 235
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 0.04,
|
| 1422 |
+
"learning_rate": 3.440237167456663e-07,
|
| 1423 |
+
"loss": 2.0745,
|
| 1424 |
+
"step": 236
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 0.04,
|
| 1428 |
+
"learning_rate": 3.337716739805264e-07,
|
| 1429 |
+
"loss": 2.0536,
|
| 1430 |
+
"step": 237
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 0.04,
|
| 1434 |
+
"learning_rate": 3.2365555427281634e-07,
|
| 1435 |
+
"loss": 2.2728,
|
| 1436 |
+
"step": 238
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 0.04,
|
| 1440 |
+
"learning_rate": 3.1367653664888173e-07,
|
| 1441 |
+
"loss": 2.1603,
|
| 1442 |
+
"step": 239
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 0.04,
|
| 1446 |
+
"learning_rate": 3.0383578415591913e-07,
|
| 1447 |
+
"loss": 2.1522,
|
| 1448 |
+
"step": 240
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 0.04,
|
| 1452 |
+
"learning_rate": 2.9413444372642496e-07,
|
| 1453 |
+
"loss": 2.2954,
|
| 1454 |
+
"step": 241
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 0.04,
|
| 1458 |
+
"learning_rate": 2.8457364604452376e-07,
|
| 1459 |
+
"loss": 2.1479,
|
| 1460 |
+
"step": 242
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 0.04,
|
| 1464 |
+
"learning_rate": 2.751545054141834e-07,
|
| 1465 |
+
"loss": 2.3303,
|
| 1466 |
+
"step": 243
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 0.04,
|
| 1470 |
+
"learning_rate": 2.6587811962934823e-07,
|
| 1471 |
+
"loss": 2.3258,
|
| 1472 |
+
"step": 244
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 0.04,
|
| 1476 |
+
"learning_rate": 2.567455698459882e-07,
|
| 1477 |
+
"loss": 2.3271,
|
| 1478 |
+
"step": 245
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 0.04,
|
| 1482 |
+
"learning_rate": 2.4775792045609353e-07,
|
| 1483 |
+
"loss": 2.3152,
|
| 1484 |
+
"step": 246
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 0.04,
|
| 1488 |
+
"learning_rate": 2.389162189636188e-07,
|
| 1489 |
+
"loss": 2.4583,
|
| 1490 |
+
"step": 247
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 0.04,
|
| 1494 |
+
"learning_rate": 2.3022149586239972e-07,
|
| 1495 |
+
"loss": 2.3737,
|
| 1496 |
+
"step": 248
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 0.04,
|
| 1500 |
+
"learning_rate": 2.2167476451604624e-07,
|
| 1501 |
+
"loss": 2.5374,
|
| 1502 |
+
"step": 249
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 0.04,
|
| 1506 |
+
"learning_rate": 2.1327702103983864e-07,
|
| 1507 |
+
"loss": 2.5436,
|
| 1508 |
+
"step": 250
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 0.04,
|
| 1512 |
+
"learning_rate": 2.0502924418463014e-07,
|
| 1513 |
+
"loss": 1.1269,
|
| 1514 |
+
"step": 251
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 0.04,
|
| 1518 |
+
"learning_rate": 1.9693239522277327e-07,
|
| 1519 |
+
"loss": 1.1588,
|
| 1520 |
+
"step": 252
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 0.04,
|
| 1524 |
+
"learning_rate": 1.8898741783608642e-07,
|
| 1525 |
+
"loss": 1.127,
|
| 1526 |
+
"step": 253
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 0.04,
|
| 1530 |
+
"learning_rate": 1.811952380058657e-07,
|
| 1531 |
+
"loss": 1.1566,
|
| 1532 |
+
"step": 254
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 0.04,
|
| 1536 |
+
"learning_rate": 1.7355676390496482e-07,
|
| 1537 |
+
"loss": 1.2204,
|
| 1538 |
+
"step": 255
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 0.04,
|
| 1542 |
+
"learning_rate": 1.660728857919464e-07,
|
| 1543 |
+
"loss": 1.1514,
|
| 1544 |
+
"step": 256
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 0.04,
|
| 1548 |
+
"learning_rate": 1.5874447590732537e-07,
|
| 1549 |
+
"loss": 1.1718,
|
| 1550 |
+
"step": 257
|
| 1551 |
+
},
|
| 1552 |
+
{
|
| 1553 |
+
"epoch": 0.04,
|
| 1554 |
+
"learning_rate": 1.5157238837190719e-07,
|
| 1555 |
+
"loss": 1.3254,
|
| 1556 |
+
"step": 258
|
| 1557 |
+
},
|
| 1558 |
+
{
|
| 1559 |
+
"epoch": 0.04,
|
| 1560 |
+
"learning_rate": 1.4455745908724226e-07,
|
| 1561 |
+
"loss": 1.2906,
|
| 1562 |
+
"step": 259
|
| 1563 |
+
},
|
| 1564 |
+
{
|
| 1565 |
+
"epoch": 0.04,
|
| 1566 |
+
"learning_rate": 1.377005056382018e-07,
|
| 1567 |
+
"loss": 1.1925,
|
| 1568 |
+
"step": 260
|
| 1569 |
+
},
|
| 1570 |
+
{
|
| 1571 |
+
"epoch": 0.04,
|
| 1572 |
+
"learning_rate": 1.3100232719768996e-07,
|
| 1573 |
+
"loss": 1.2717,
|
| 1574 |
+
"step": 261
|
| 1575 |
+
},
|
| 1576 |
+
{
|
| 1577 |
+
"epoch": 0.04,
|
| 1578 |
+
"learning_rate": 1.2446370443349863e-07,
|
| 1579 |
+
"loss": 1.2414,
|
| 1580 |
+
"step": 262
|
| 1581 |
+
},
|
| 1582 |
+
{
|
| 1583 |
+
"epoch": 0.04,
|
| 1584 |
+
"learning_rate": 1.180853994173236e-07,
|
| 1585 |
+
"loss": 1.2571,
|
| 1586 |
+
"step": 263
|
| 1587 |
+
},
|
| 1588 |
+
{
|
| 1589 |
+
"epoch": 0.04,
|
| 1590 |
+
"learning_rate": 1.1186815553594382e-07,
|
| 1591 |
+
"loss": 1.1211,
|
| 1592 |
+
"step": 264
|
| 1593 |
+
},
|
| 1594 |
+
{
|
| 1595 |
+
"epoch": 0.04,
|
| 1596 |
+
"learning_rate": 1.058126974045811e-07,
|
| 1597 |
+
"loss": 1.3869,
|
| 1598 |
+
"step": 265
|
| 1599 |
+
},
|
| 1600 |
+
{
|
| 1601 |
+
"epoch": 0.04,
|
| 1602 |
+
"learning_rate": 9.991973078244638e-08,
|
| 1603 |
+
"loss": 1.4214,
|
| 1604 |
+
"step": 266
|
| 1605 |
+
},
|
| 1606 |
+
{
|
| 1607 |
+
"epoch": 0.04,
|
| 1608 |
+
"learning_rate": 9.418994249048474e-08,
|
| 1609 |
+
"loss": 1.4063,
|
| 1610 |
+
"step": 267
|
| 1611 |
+
},
|
| 1612 |
+
{
|
| 1613 |
+
"epoch": 0.04,
|
| 1614 |
+
"learning_rate": 8.862400033132573e-08,
|
| 1615 |
+
"loss": 1.4621,
|
| 1616 |
+
"step": 268
|
| 1617 |
+
},
|
| 1618 |
+
{
|
| 1619 |
+
"epoch": 0.04,
|
| 1620 |
+
"learning_rate": 8.322255301145204e-08,
|
| 1621 |
+
"loss": 1.5625,
|
| 1622 |
+
"step": 269
|
| 1623 |
+
},
|
| 1624 |
+
{
|
| 1625 |
+
"epoch": 0.04,
|
| 1626 |
+
"learning_rate": 7.798623006559436e-08,
|
| 1627 |
+
"loss": 1.519,
|
| 1628 |
+
"step": 270
|
| 1629 |
+
},
|
| 1630 |
+
{
|
| 1631 |
+
"epoch": 0.04,
|
| 1632 |
+
"learning_rate": 7.291564178335719e-08,
|
| 1633 |
+
"loss": 1.5473,
|
| 1634 |
+
"step": 271
|
| 1635 |
+
},
|
| 1636 |
+
{
|
| 1637 |
+
"epoch": 0.04,
|
| 1638 |
+
"learning_rate": 6.801137913809214e-08,
|
| 1639 |
+
"loss": 1.5158,
|
| 1640 |
+
"step": 272
|
| 1641 |
+
},
|
| 1642 |
+
{
|
| 1643 |
+
"epoch": 0.04,
|
| 1644 |
+
"learning_rate": 6.327401371801944e-08,
|
| 1645 |
+
"loss": 1.6969,
|
| 1646 |
+
"step": 273
|
| 1647 |
+
},
|
| 1648 |
+
{
|
| 1649 |
+
"epoch": 0.04,
|
| 1650 |
+
"learning_rate": 5.870409765960966e-08,
|
| 1651 |
+
"loss": 1.6519,
|
| 1652 |
+
"step": 274
|
| 1653 |
+
},
|
| 1654 |
+
{
|
| 1655 |
+
"epoch": 0.04,
|
| 1656 |
+
"learning_rate": 5.430216358323309e-08,
|
| 1657 |
+
"loss": 1.5961,
|
| 1658 |
+
"step": 275
|
| 1659 |
+
},
|
| 1660 |
+
{
|
| 1661 |
+
"epoch": 0.04,
|
| 1662 |
+
"learning_rate": 5.00687245310833e-08,
|
| 1663 |
+
"loss": 1.6128,
|
| 1664 |
+
"step": 276
|
| 1665 |
+
},
|
| 1666 |
+
{
|
| 1667 |
+
"epoch": 0.04,
|
| 1668 |
+
"learning_rate": 4.60042739073816e-08,
|
| 1669 |
+
"loss": 1.813,
|
| 1670 |
+
"step": 277
|
| 1671 |
+
},
|
| 1672 |
+
{
|
| 1673 |
+
"epoch": 0.04,
|
| 1674 |
+
"learning_rate": 4.2109285420872055e-08,
|
| 1675 |
+
"loss": 1.6032,
|
| 1676 |
+
"step": 278
|
| 1677 |
+
},
|
| 1678 |
+
{
|
| 1679 |
+
"epoch": 0.04,
|
| 1680 |
+
"learning_rate": 3.838421302961098e-08,
|
| 1681 |
+
"loss": 1.8927,
|
| 1682 |
+
"step": 279
|
| 1683 |
+
},
|
| 1684 |
+
{
|
| 1685 |
+
"epoch": 0.04,
|
| 1686 |
+
"learning_rate": 3.4829490888057424e-08,
|
| 1687 |
+
"loss": 1.9264,
|
| 1688 |
+
"step": 280
|
| 1689 |
+
},
|
| 1690 |
+
{
|
| 1691 |
+
"epoch": 0.04,
|
| 1692 |
+
"learning_rate": 3.1445533296474484e-08,
|
| 1693 |
+
"loss": 1.7726,
|
| 1694 |
+
"step": 281
|
| 1695 |
+
},
|
| 1696 |
+
{
|
| 1697 |
+
"epoch": 0.04,
|
| 1698 |
+
"learning_rate": 2.8232734652641424e-08,
|
| 1699 |
+
"loss": 1.9217,
|
| 1700 |
+
"step": 282
|
| 1701 |
+
},
|
| 1702 |
+
{
|
| 1703 |
+
"epoch": 0.04,
|
| 1704 |
+
"learning_rate": 2.5191469405887625e-08,
|
| 1705 |
+
"loss": 1.9583,
|
| 1706 |
+
"step": 283
|
| 1707 |
+
},
|
| 1708 |
+
{
|
| 1709 |
+
"epoch": 0.04,
|
| 1710 |
+
"learning_rate": 2.2322092013450313e-08,
|
| 1711 |
+
"loss": 2.0714,
|
| 1712 |
+
"step": 284
|
| 1713 |
+
},
|
| 1714 |
+
{
|
| 1715 |
+
"epoch": 0.04,
|
| 1716 |
+
"learning_rate": 1.962493689916395e-08,
|
| 1717 |
+
"loss": 2.0708,
|
| 1718 |
+
"step": 285
|
| 1719 |
+
},
|
| 1720 |
+
{
|
| 1721 |
+
"epoch": 0.04,
|
| 1722 |
+
"learning_rate": 1.7100318414482063e-08,
|
| 1723 |
+
"loss": 2.0683,
|
| 1724 |
+
"step": 286
|
| 1725 |
+
},
|
| 1726 |
+
{
|
| 1727 |
+
"epoch": 0.04,
|
| 1728 |
+
"learning_rate": 1.4748530801840076e-08,
|
| 1729 |
+
"loss": 2.1105,
|
| 1730 |
+
"step": 287
|
| 1731 |
+
},
|
| 1732 |
+
{
|
| 1733 |
+
"epoch": 0.04,
|
| 1734 |
+
"learning_rate": 1.2569848160362384e-08,
|
| 1735 |
+
"loss": 2.2285,
|
| 1736 |
+
"step": 288
|
| 1737 |
+
},
|
| 1738 |
+
{
|
| 1739 |
+
"epoch": 0.04,
|
| 1740 |
+
"learning_rate": 1.0564524413915422e-08,
|
| 1741 |
+
"loss": 2.288,
|
| 1742 |
+
"step": 289
|
| 1743 |
+
},
|
| 1744 |
+
{
|
| 1745 |
+
"epoch": 0.04,
|
| 1746 |
+
"learning_rate": 8.732793281513663e-09,
|
| 1747 |
+
"loss": 2.2546,
|
| 1748 |
+
"step": 290
|
| 1749 |
+
},
|
| 1750 |
+
{
|
| 1751 |
+
"epoch": 0.04,
|
| 1752 |
+
"learning_rate": 7.074868250079081e-09,
|
| 1753 |
+
"loss": 2.156,
|
| 1754 |
+
"step": 291
|
| 1755 |
+
},
|
| 1756 |
+
{
|
| 1757 |
+
"epoch": 0.05,
|
| 1758 |
+
"learning_rate": 5.590942549560052e-09,
|
| 1759 |
+
"loss": 2.2368,
|
| 1760 |
+
"step": 292
|
| 1761 |
+
},
|
| 1762 |
+
{
|
| 1763 |
+
"epoch": 0.05,
|
| 1764 |
+
"learning_rate": 4.2811891304105345e-09,
|
| 1765 |
+
"loss": 2.2568,
|
| 1766 |
+
"step": 293
|
| 1767 |
+
},
|
| 1768 |
+
{
|
| 1769 |
+
"epoch": 0.05,
|
| 1770 |
+
"learning_rate": 3.145760643432527e-09,
|
| 1771 |
+
"loss": 2.2339,
|
| 1772 |
+
"step": 294
|
| 1773 |
+
},
|
| 1774 |
+
{
|
| 1775 |
+
"epoch": 0.05,
|
| 1776 |
+
"learning_rate": 2.1847894219846343e-09,
|
| 1777 |
+
"loss": 2.3651,
|
| 1778 |
+
"step": 295
|
| 1779 |
+
},
|
| 1780 |
+
{
|
| 1781 |
+
"epoch": 0.05,
|
| 1782 |
+
"learning_rate": 1.3983874665589035e-09,
|
| 1783 |
+
"loss": 2.492,
|
| 1784 |
+
"step": 296
|
| 1785 |
+
},
|
| 1786 |
+
{
|
| 1787 |
+
"epoch": 0.05,
|
| 1788 |
+
"learning_rate": 7.866464317276001e-10,
|
| 1789 |
+
"loss": 2.3863,
|
| 1790 |
+
"step": 297
|
| 1791 |
+
},
|
| 1792 |
+
{
|
| 1793 |
+
"epoch": 0.05,
|
| 1794 |
+
"learning_rate": 3.496376154604186e-10,
|
| 1795 |
+
"loss": 2.4482,
|
| 1796 |
+
"step": 298
|
| 1797 |
+
},
|
| 1798 |
+
{
|
| 1799 |
+
"epoch": 0.05,
|
| 1800 |
+
"learning_rate": 8.741195081479747e-11,
|
| 1801 |
+
"loss": 2.5423,
|
| 1802 |
+
"step": 299
|
| 1803 |
+
},
|
| 1804 |
+
{
|
| 1805 |
+
"epoch": 0.05,
|
| 1806 |
+
"learning_rate": 0.0,
|
| 1807 |
+
"loss": 2.5864,
|
| 1808 |
+
"step": 300
|
| 1809 |
+
}
|
| 1810 |
+
],
|
| 1811 |
+
"max_steps": 300,
|
| 1812 |
+
"num_train_epochs": 1,
|
| 1813 |
+
"total_flos": 1.174396008923136e+16,
|
| 1814 |
+
"trial_name": null,
|
| 1815 |
+
"trial_params": null
|
| 1816 |
+
}
|
checkpoint-300/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4373b7353f7618ba071e8149528a0ae191779985a891ab40089b4f1cda12cac9
|
| 3 |
+
size 4027
|
flyte_training_config.json
CHANGED
|
@@ -1 +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":
|
|
|
|
| 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}
|
training_args.bin
CHANGED
|
@@ -1,3 +1,3 @@
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
-
oid sha256:
|
| 3 |
-
size
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4373b7353f7618ba071e8149528a0ae191779985a891ab40089b4f1cda12cac9
|
| 3 |
+
size 4027
|