Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,5 +1,5 @@
|
|
| 1 |
import gradio as gr
|
| 2 |
-
from transformers import AutoModelForCausalLM, AutoTokenizer, Trainer, TrainingArguments
|
| 3 |
from datasets import load_dataset, Dataset, DatasetDict
|
| 4 |
import os
|
| 5 |
import time
|
|
@@ -107,7 +107,7 @@ class CustomCallback(TrainerCallback):
|
|
| 107 |
|
| 108 |
def on_step_begin(self, args, state, control, **kwargs):
|
| 109 |
global progress_info
|
| 110 |
-
total_steps = state.
|
| 111 |
current_step = state.global_step
|
| 112 |
progress_info["status"] = f"銈ㄣ儩銉冦偗 {state.epoch + 1} / {args.num_train_epochs}, 銈广儐銉冦儣 {current_step + 1} / {total_steps}"
|
| 113 |
progress_info["progress"] = (current_step + 1) / total_steps
|
|
@@ -115,7 +115,7 @@ class CustomCallback(TrainerCallback):
|
|
| 115 |
|
| 116 |
def on_step_end(self, args, state, control, **kwargs):
|
| 117 |
global progress_info
|
| 118 |
-
total_steps = state.
|
| 119 |
current_step = state.global_step
|
| 120 |
elapsed_time = time.time() - state.log_history[0]["epoch_time"]
|
| 121 |
time_per_step = elapsed_time / (current_step + 1)
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer, Trainer, TrainingArguments, TrainerCallback
|
| 3 |
from datasets import load_dataset, Dataset, DatasetDict
|
| 4 |
import os
|
| 5 |
import time
|
|
|
|
| 107 |
|
| 108 |
def on_step_begin(self, args, state, control, **kwargs):
|
| 109 |
global progress_info
|
| 110 |
+
total_steps = state.max_steps
|
| 111 |
current_step = state.global_step
|
| 112 |
progress_info["status"] = f"銈ㄣ儩銉冦偗 {state.epoch + 1} / {args.num_train_epochs}, 銈广儐銉冦儣 {current_step + 1} / {total_steps}"
|
| 113 |
progress_info["progress"] = (current_step + 1) / total_steps
|
|
|
|
| 115 |
|
| 116 |
def on_step_end(self, args, state, control, **kwargs):
|
| 117 |
global progress_info
|
| 118 |
+
total_steps = state.max_steps
|
| 119 |
current_step = state.global_step
|
| 120 |
elapsed_time = time.time() - state.log_history[0]["epoch_time"]
|
| 121 |
time_per_step = elapsed_time / (current_step + 1)
|