Add app and some data to model
Browse files- app.py +166 -0
- chkp/adapter_config.json +20 -0
- chkp/adapter_model.bin +3 -0
- chkp/optimizer.pt +3 -0
- chkp/rng_state.pth +3 -0
- chkp/scheduler.pt +3 -0
- chkp/trainer_state.json +349 -0
- chkp/training_args.bin +3 -0
app.py
ADDED
|
@@ -0,0 +1,166 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import sys
|
| 3 |
+
|
| 4 |
+
import fire
|
| 5 |
+
import gradio as gr
|
| 6 |
+
import torch
|
| 7 |
+
import transformers
|
| 8 |
+
from peft import PeftModel
|
| 9 |
+
from transformers import GenerationConfig, LlamaForCausalLM, LlamaTokenizer
|
| 10 |
+
|
| 11 |
+
from typing import Union
|
| 12 |
+
import re
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
class Prompter(object):
|
| 16 |
+
def generate_prompt(
|
| 17 |
+
self,
|
| 18 |
+
instruction: str,
|
| 19 |
+
label: Union[None, str] = None,
|
| 20 |
+
) -> str:
|
| 21 |
+
res = f"{instruction}\nAnswer: "
|
| 22 |
+
|
| 23 |
+
if label:
|
| 24 |
+
res = f"{res}{label}"
|
| 25 |
+
|
| 26 |
+
return res
|
| 27 |
+
|
| 28 |
+
def get_response(self, output: str) -> str:
|
| 29 |
+
return (
|
| 30 |
+
output.split("Answer:")[1]
|
| 31 |
+
.strip()
|
| 32 |
+
.replace("/", "\u00F7")
|
| 33 |
+
.replace("*", "\u00D7")
|
| 34 |
+
)
|
| 35 |
+
|
| 36 |
+
|
| 37 |
+
load_8bit = True # for Colab
|
| 38 |
+
base_model = "baffo32/decapoda-research-llama-7B-hf"
|
| 39 |
+
lora_weights = "tiedong/goat-lora-7b"
|
| 40 |
+
share_gradio = True
|
| 41 |
+
|
| 42 |
+
if torch.cuda.is_available():
|
| 43 |
+
device = "cuda"
|
| 44 |
+
else:
|
| 45 |
+
device = "cpu"
|
| 46 |
+
|
| 47 |
+
try:
|
| 48 |
+
if torch.backends.mps.is_available():
|
| 49 |
+
device = "mps"
|
| 50 |
+
except:
|
| 51 |
+
pass
|
| 52 |
+
|
| 53 |
+
prompter = Prompter()
|
| 54 |
+
|
| 55 |
+
tokenizer = LlamaTokenizer.from_pretrained("hf-internal-testing/llama-tokenizer")
|
| 56 |
+
if device == "cuda":
|
| 57 |
+
model = LlamaForCausalLM.from_pretrained(
|
| 58 |
+
base_model,
|
| 59 |
+
load_in_8bit=load_8bit,
|
| 60 |
+
torch_dtype=torch.float16,
|
| 61 |
+
device_map="auto",
|
| 62 |
+
)
|
| 63 |
+
model = PeftModel.from_pretrained(
|
| 64 |
+
model,
|
| 65 |
+
lora_weights,
|
| 66 |
+
torch_dtype=torch.float16,
|
| 67 |
+
device_map={"": 0},
|
| 68 |
+
)
|
| 69 |
+
elif device == "mps":
|
| 70 |
+
model = LlamaForCausalLM.from_pretrained(
|
| 71 |
+
base_model,
|
| 72 |
+
device_map={"": device},
|
| 73 |
+
torch_dtype=torch.float16,
|
| 74 |
+
)
|
| 75 |
+
model = PeftModel.from_pretrained(
|
| 76 |
+
model,
|
| 77 |
+
lora_weights,
|
| 78 |
+
device_map={"": device},
|
| 79 |
+
torch_dtype=torch.float16,
|
| 80 |
+
)
|
| 81 |
+
else:
|
| 82 |
+
model = LlamaForCausalLM.from_pretrained(
|
| 83 |
+
base_model, device_map={"": device}, low_cpu_mem_usage=True
|
| 84 |
+
)
|
| 85 |
+
model = PeftModel.from_pretrained(
|
| 86 |
+
model,
|
| 87 |
+
lora_weights,
|
| 88 |
+
device_map={"": device},
|
| 89 |
+
)
|
| 90 |
+
|
| 91 |
+
if not load_8bit:
|
| 92 |
+
model.half()
|
| 93 |
+
|
| 94 |
+
model.eval()
|
| 95 |
+
if torch.__version__ >= "2" and sys.platform != "win32":
|
| 96 |
+
model = torch.compile(model)
|
| 97 |
+
|
| 98 |
+
|
| 99 |
+
def evaluate(
|
| 100 |
+
instruction,
|
| 101 |
+
temperature=0.1,
|
| 102 |
+
top_p=0.75,
|
| 103 |
+
top_k=40,
|
| 104 |
+
num_beams=4,
|
| 105 |
+
max_new_tokens=512,
|
| 106 |
+
stream_output=True,
|
| 107 |
+
**kwargs,
|
| 108 |
+
):
|
| 109 |
+
prompt = prompter.generate_prompt(instruction)
|
| 110 |
+
inputs = tokenizer(prompt, return_tensors="pt")
|
| 111 |
+
input_ids = inputs["input_ids"].to(device)
|
| 112 |
+
generation_config = GenerationConfig(
|
| 113 |
+
temperature=temperature,
|
| 114 |
+
top_p=top_p,
|
| 115 |
+
top_k=top_k,
|
| 116 |
+
num_beams=num_beams,
|
| 117 |
+
**kwargs,
|
| 118 |
+
)
|
| 119 |
+
|
| 120 |
+
generate_params = {
|
| 121 |
+
"input_ids": input_ids,
|
| 122 |
+
"generation_config": generation_config,
|
| 123 |
+
"return_dict_in_generate": True,
|
| 124 |
+
"output_scores": True,
|
| 125 |
+
"max_new_tokens": max_new_tokens,
|
| 126 |
+
}
|
| 127 |
+
|
| 128 |
+
# Without streaming
|
| 129 |
+
with torch.no_grad():
|
| 130 |
+
generation_output = model.generate(
|
| 131 |
+
input_ids=input_ids,
|
| 132 |
+
generation_config=generation_config,
|
| 133 |
+
return_dict_in_generate=True,
|
| 134 |
+
output_scores=True,
|
| 135 |
+
max_new_tokens=max_new_tokens,
|
| 136 |
+
)
|
| 137 |
+
s = generation_output.sequences[0]
|
| 138 |
+
output = tokenizer.decode(s, skip_special_tokens=True).strip()
|
| 139 |
+
yield prompter.get_response(output)
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
gr.Interface(
|
| 143 |
+
fn=evaluate,
|
| 144 |
+
inputs=[
|
| 145 |
+
gr.components.Textbox(
|
| 146 |
+
lines=1,
|
| 147 |
+
label="Arithmetic",
|
| 148 |
+
placeholder="What is 63303235 + 20239503",
|
| 149 |
+
),
|
| 150 |
+
gr.components.Slider(minimum=0, maximum=1, value=0.1, label="Temperature"),
|
| 151 |
+
gr.components.Slider(minimum=0, maximum=1, value=0.75, label="Top p"),
|
| 152 |
+
gr.components.Slider(minimum=0, maximum=100, step=1, value=40, label="Top k"),
|
| 153 |
+
gr.components.Slider(minimum=1, maximum=4, step=1, value=4, label="Beams"),
|
| 154 |
+
gr.components.Slider(
|
| 155 |
+
minimum=1, maximum=1024, step=1, value=512, label="Max tokens"
|
| 156 |
+
),
|
| 157 |
+
],
|
| 158 |
+
outputs=[
|
| 159 |
+
gr.Textbox(
|
| 160 |
+
lines=5,
|
| 161 |
+
label="Output",
|
| 162 |
+
)
|
| 163 |
+
],
|
| 164 |
+
title="test model",
|
| 165 |
+
description="Это пример реализации из goat", # noqa: E501
|
| 166 |
+
).queue().launch(share=share_gradio)
|
chkp/adapter_config.json
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"base_model_name_or_path": "nickypro/tinyllama-15M",
|
| 3 |
+
"bias": "none",
|
| 4 |
+
"enable_lora": null,
|
| 5 |
+
"fan_in_fan_out": false,
|
| 6 |
+
"inference_mode": true,
|
| 7 |
+
"lora_alpha": 64,
|
| 8 |
+
"lora_dropout": 0.05,
|
| 9 |
+
"merge_weights": false,
|
| 10 |
+
"modules_to_save": null,
|
| 11 |
+
"peft_type": "LORA",
|
| 12 |
+
"r": 64,
|
| 13 |
+
"target_modules": [
|
| 14 |
+
"q_proj",
|
| 15 |
+
"v_proj",
|
| 16 |
+
"k_proj",
|
| 17 |
+
"o_proj"
|
| 18 |
+
],
|
| 19 |
+
"task_type": "CAUSAL_LM"
|
| 20 |
+
}
|
chkp/adapter_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:49891d6a9e5d6098f4048189ae2bf4df53022b58c7689f25da4e6b9c481a018a
|
| 3 |
+
size 3556350
|
chkp/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e68114a0c0afc9a77f9bd91ed114283dc2093c3bf42383e62897601f2b4f8129
|
| 3 |
+
size 7118586
|
chkp/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cb69e610873a8142e0245cca374768d45ddb50a0c4891436e6dbf04d069a7122
|
| 3 |
+
size 14244
|
chkp/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:90ba5da359c992d8f0dc912ca51bea8184b55a457a5f5fdfc7bc2702765e7d0d
|
| 3 |
+
size 1064
|
chkp/trainer_state.json
ADDED
|
@@ -0,0 +1,349 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": null,
|
| 3 |
+
"best_model_checkpoint": null,
|
| 4 |
+
"epoch": 0.23157894736842105,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 550,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.0,
|
| 13 |
+
"learning_rate": 1e-05,
|
| 14 |
+
"loss": 4.8744,
|
| 15 |
+
"step": 10
|
| 16 |
+
},
|
| 17 |
+
{
|
| 18 |
+
"epoch": 0.01,
|
| 19 |
+
"learning_rate": 2e-05,
|
| 20 |
+
"loss": 4.1114,
|
| 21 |
+
"step": 20
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"epoch": 0.01,
|
| 25 |
+
"learning_rate": 3e-05,
|
| 26 |
+
"loss": 3.528,
|
| 27 |
+
"step": 30
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"epoch": 0.02,
|
| 31 |
+
"learning_rate": 4e-05,
|
| 32 |
+
"loss": 3.2573,
|
| 33 |
+
"step": 40
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 0.02,
|
| 37 |
+
"learning_rate": 5e-05,
|
| 38 |
+
"loss": 3.1417,
|
| 39 |
+
"step": 50
|
| 40 |
+
},
|
| 41 |
+
{
|
| 42 |
+
"epoch": 0.03,
|
| 43 |
+
"learning_rate": 6e-05,
|
| 44 |
+
"loss": 3.0506,
|
| 45 |
+
"step": 60
|
| 46 |
+
},
|
| 47 |
+
{
|
| 48 |
+
"epoch": 0.03,
|
| 49 |
+
"learning_rate": 7e-05,
|
| 50 |
+
"loss": 2.9644,
|
| 51 |
+
"step": 70
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.03,
|
| 55 |
+
"learning_rate": 8e-05,
|
| 56 |
+
"loss": 2.8399,
|
| 57 |
+
"step": 80
|
| 58 |
+
},
|
| 59 |
+
{
|
| 60 |
+
"epoch": 0.04,
|
| 61 |
+
"learning_rate": 9e-05,
|
| 62 |
+
"loss": 2.6935,
|
| 63 |
+
"step": 90
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 0.04,
|
| 67 |
+
"learning_rate": 0.0001,
|
| 68 |
+
"loss": 2.5259,
|
| 69 |
+
"step": 100
|
| 70 |
+
},
|
| 71 |
+
{
|
| 72 |
+
"epoch": 0.05,
|
| 73 |
+
"learning_rate": 9.956043956043956e-05,
|
| 74 |
+
"loss": 2.3713,
|
| 75 |
+
"step": 110
|
| 76 |
+
},
|
| 77 |
+
{
|
| 78 |
+
"epoch": 0.05,
|
| 79 |
+
"learning_rate": 9.912087912087913e-05,
|
| 80 |
+
"loss": 2.2237,
|
| 81 |
+
"step": 120
|
| 82 |
+
},
|
| 83 |
+
{
|
| 84 |
+
"epoch": 0.05,
|
| 85 |
+
"learning_rate": 9.868131868131869e-05,
|
| 86 |
+
"loss": 2.0999,
|
| 87 |
+
"step": 130
|
| 88 |
+
},
|
| 89 |
+
{
|
| 90 |
+
"epoch": 0.06,
|
| 91 |
+
"learning_rate": 9.824175824175824e-05,
|
| 92 |
+
"loss": 2.0211,
|
| 93 |
+
"step": 140
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.06,
|
| 97 |
+
"learning_rate": 9.780219780219781e-05,
|
| 98 |
+
"loss": 1.949,
|
| 99 |
+
"step": 150
|
| 100 |
+
},
|
| 101 |
+
{
|
| 102 |
+
"epoch": 0.07,
|
| 103 |
+
"learning_rate": 9.736263736263737e-05,
|
| 104 |
+
"loss": 1.8819,
|
| 105 |
+
"step": 160
|
| 106 |
+
},
|
| 107 |
+
{
|
| 108 |
+
"epoch": 0.07,
|
| 109 |
+
"learning_rate": 9.692307692307692e-05,
|
| 110 |
+
"loss": 1.8244,
|
| 111 |
+
"step": 170
|
| 112 |
+
},
|
| 113 |
+
{
|
| 114 |
+
"epoch": 0.08,
|
| 115 |
+
"learning_rate": 9.64835164835165e-05,
|
| 116 |
+
"loss": 1.7849,
|
| 117 |
+
"step": 180
|
| 118 |
+
},
|
| 119 |
+
{
|
| 120 |
+
"epoch": 0.08,
|
| 121 |
+
"learning_rate": 9.604395604395605e-05,
|
| 122 |
+
"loss": 1.7288,
|
| 123 |
+
"step": 190
|
| 124 |
+
},
|
| 125 |
+
{
|
| 126 |
+
"epoch": 0.08,
|
| 127 |
+
"learning_rate": 9.560439560439561e-05,
|
| 128 |
+
"loss": 1.6914,
|
| 129 |
+
"step": 200
|
| 130 |
+
},
|
| 131 |
+
{
|
| 132 |
+
"epoch": 0.09,
|
| 133 |
+
"learning_rate": 9.516483516483517e-05,
|
| 134 |
+
"loss": 1.6534,
|
| 135 |
+
"step": 210
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.09,
|
| 139 |
+
"learning_rate": 9.472527472527473e-05,
|
| 140 |
+
"loss": 1.6332,
|
| 141 |
+
"step": 220
|
| 142 |
+
},
|
| 143 |
+
{
|
| 144 |
+
"epoch": 0.1,
|
| 145 |
+
"learning_rate": 9.428571428571429e-05,
|
| 146 |
+
"loss": 1.5944,
|
| 147 |
+
"step": 230
|
| 148 |
+
},
|
| 149 |
+
{
|
| 150 |
+
"epoch": 0.1,
|
| 151 |
+
"learning_rate": 9.384615384615386e-05,
|
| 152 |
+
"loss": 1.5789,
|
| 153 |
+
"step": 240
|
| 154 |
+
},
|
| 155 |
+
{
|
| 156 |
+
"epoch": 0.11,
|
| 157 |
+
"learning_rate": 9.340659340659341e-05,
|
| 158 |
+
"loss": 1.5583,
|
| 159 |
+
"step": 250
|
| 160 |
+
},
|
| 161 |
+
{
|
| 162 |
+
"epoch": 0.11,
|
| 163 |
+
"learning_rate": 9.296703296703297e-05,
|
| 164 |
+
"loss": 1.5485,
|
| 165 |
+
"step": 260
|
| 166 |
+
},
|
| 167 |
+
{
|
| 168 |
+
"epoch": 0.11,
|
| 169 |
+
"learning_rate": 9.252747252747253e-05,
|
| 170 |
+
"loss": 1.5124,
|
| 171 |
+
"step": 270
|
| 172 |
+
},
|
| 173 |
+
{
|
| 174 |
+
"epoch": 0.12,
|
| 175 |
+
"learning_rate": 9.208791208791209e-05,
|
| 176 |
+
"loss": 1.5004,
|
| 177 |
+
"step": 280
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.12,
|
| 181 |
+
"learning_rate": 9.164835164835165e-05,
|
| 182 |
+
"loss": 1.5042,
|
| 183 |
+
"step": 290
|
| 184 |
+
},
|
| 185 |
+
{
|
| 186 |
+
"epoch": 0.13,
|
| 187 |
+
"learning_rate": 9.12087912087912e-05,
|
| 188 |
+
"loss": 1.4991,
|
| 189 |
+
"step": 300
|
| 190 |
+
},
|
| 191 |
+
{
|
| 192 |
+
"epoch": 0.13,
|
| 193 |
+
"learning_rate": 9.076923076923078e-05,
|
| 194 |
+
"loss": 1.4692,
|
| 195 |
+
"step": 310
|
| 196 |
+
},
|
| 197 |
+
{
|
| 198 |
+
"epoch": 0.13,
|
| 199 |
+
"learning_rate": 9.032967032967033e-05,
|
| 200 |
+
"loss": 1.4628,
|
| 201 |
+
"step": 320
|
| 202 |
+
},
|
| 203 |
+
{
|
| 204 |
+
"epoch": 0.14,
|
| 205 |
+
"learning_rate": 8.989010989010989e-05,
|
| 206 |
+
"loss": 1.4481,
|
| 207 |
+
"step": 330
|
| 208 |
+
},
|
| 209 |
+
{
|
| 210 |
+
"epoch": 0.14,
|
| 211 |
+
"learning_rate": 8.945054945054946e-05,
|
| 212 |
+
"loss": 1.454,
|
| 213 |
+
"step": 340
|
| 214 |
+
},
|
| 215 |
+
{
|
| 216 |
+
"epoch": 0.15,
|
| 217 |
+
"learning_rate": 8.901098901098901e-05,
|
| 218 |
+
"loss": 1.4409,
|
| 219 |
+
"step": 350
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.15,
|
| 223 |
+
"learning_rate": 8.857142857142857e-05,
|
| 224 |
+
"loss": 1.4339,
|
| 225 |
+
"step": 360
|
| 226 |
+
},
|
| 227 |
+
{
|
| 228 |
+
"epoch": 0.16,
|
| 229 |
+
"learning_rate": 8.813186813186814e-05,
|
| 230 |
+
"loss": 1.4262,
|
| 231 |
+
"step": 370
|
| 232 |
+
},
|
| 233 |
+
{
|
| 234 |
+
"epoch": 0.16,
|
| 235 |
+
"learning_rate": 8.76923076923077e-05,
|
| 236 |
+
"loss": 1.4298,
|
| 237 |
+
"step": 380
|
| 238 |
+
},
|
| 239 |
+
{
|
| 240 |
+
"epoch": 0.16,
|
| 241 |
+
"learning_rate": 8.725274725274725e-05,
|
| 242 |
+
"loss": 1.4247,
|
| 243 |
+
"step": 390
|
| 244 |
+
},
|
| 245 |
+
{
|
| 246 |
+
"epoch": 0.17,
|
| 247 |
+
"learning_rate": 8.681318681318682e-05,
|
| 248 |
+
"loss": 1.4191,
|
| 249 |
+
"step": 400
|
| 250 |
+
},
|
| 251 |
+
{
|
| 252 |
+
"epoch": 0.17,
|
| 253 |
+
"learning_rate": 8.637362637362638e-05,
|
| 254 |
+
"loss": 1.4156,
|
| 255 |
+
"step": 410
|
| 256 |
+
},
|
| 257 |
+
{
|
| 258 |
+
"epoch": 0.18,
|
| 259 |
+
"learning_rate": 8.593406593406593e-05,
|
| 260 |
+
"loss": 1.4102,
|
| 261 |
+
"step": 420
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.18,
|
| 265 |
+
"learning_rate": 8.54945054945055e-05,
|
| 266 |
+
"loss": 1.4121,
|
| 267 |
+
"step": 430
|
| 268 |
+
},
|
| 269 |
+
{
|
| 270 |
+
"epoch": 0.19,
|
| 271 |
+
"learning_rate": 8.505494505494506e-05,
|
| 272 |
+
"loss": 1.4033,
|
| 273 |
+
"step": 440
|
| 274 |
+
},
|
| 275 |
+
{
|
| 276 |
+
"epoch": 0.19,
|
| 277 |
+
"learning_rate": 8.461538461538461e-05,
|
| 278 |
+
"loss": 1.3986,
|
| 279 |
+
"step": 450
|
| 280 |
+
},
|
| 281 |
+
{
|
| 282 |
+
"epoch": 0.19,
|
| 283 |
+
"learning_rate": 8.417582417582419e-05,
|
| 284 |
+
"loss": 1.3935,
|
| 285 |
+
"step": 460
|
| 286 |
+
},
|
| 287 |
+
{
|
| 288 |
+
"epoch": 0.2,
|
| 289 |
+
"learning_rate": 8.373626373626374e-05,
|
| 290 |
+
"loss": 1.399,
|
| 291 |
+
"step": 470
|
| 292 |
+
},
|
| 293 |
+
{
|
| 294 |
+
"epoch": 0.2,
|
| 295 |
+
"learning_rate": 8.32967032967033e-05,
|
| 296 |
+
"loss": 1.3888,
|
| 297 |
+
"step": 480
|
| 298 |
+
},
|
| 299 |
+
{
|
| 300 |
+
"epoch": 0.21,
|
| 301 |
+
"learning_rate": 8.285714285714287e-05,
|
| 302 |
+
"loss": 1.3958,
|
| 303 |
+
"step": 490
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.21,
|
| 307 |
+
"learning_rate": 8.241758241758242e-05,
|
| 308 |
+
"loss": 1.3758,
|
| 309 |
+
"step": 500
|
| 310 |
+
},
|
| 311 |
+
{
|
| 312 |
+
"epoch": 0.21,
|
| 313 |
+
"learning_rate": 8.197802197802198e-05,
|
| 314 |
+
"loss": 1.3779,
|
| 315 |
+
"step": 510
|
| 316 |
+
},
|
| 317 |
+
{
|
| 318 |
+
"epoch": 0.22,
|
| 319 |
+
"learning_rate": 8.153846153846155e-05,
|
| 320 |
+
"loss": 1.3754,
|
| 321 |
+
"step": 520
|
| 322 |
+
},
|
| 323 |
+
{
|
| 324 |
+
"epoch": 0.22,
|
| 325 |
+
"learning_rate": 8.10989010989011e-05,
|
| 326 |
+
"loss": 1.3754,
|
| 327 |
+
"step": 530
|
| 328 |
+
},
|
| 329 |
+
{
|
| 330 |
+
"epoch": 0.23,
|
| 331 |
+
"learning_rate": 8.065934065934066e-05,
|
| 332 |
+
"loss": 1.3854,
|
| 333 |
+
"step": 540
|
| 334 |
+
},
|
| 335 |
+
{
|
| 336 |
+
"epoch": 0.23,
|
| 337 |
+
"learning_rate": 8.021978021978022e-05,
|
| 338 |
+
"loss": 1.3579,
|
| 339 |
+
"step": 550
|
| 340 |
+
}
|
| 341 |
+
],
|
| 342 |
+
"logging_steps": 10,
|
| 343 |
+
"max_steps": 2375,
|
| 344 |
+
"num_train_epochs": 1,
|
| 345 |
+
"save_steps": 50,
|
| 346 |
+
"total_flos": 271592074027008.0,
|
| 347 |
+
"trial_name": null,
|
| 348 |
+
"trial_params": null
|
| 349 |
+
}
|
chkp/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:6f2106e8a7b7aaf8d6c8f70689f1a1f3239d79f9126f142a9ce1a3ba53c11449
|
| 3 |
+
size 4536
|