Spaces:
Running
on
Zero
Running
on
Zero
remove BetterTransformer for classifier
Browse filesalready integrated into transformers based on error
app.py
CHANGED
|
@@ -1,38 +1,30 @@
|
|
| 1 |
-
import re
|
| 2 |
-
import os
|
| 3 |
import gc
|
|
|
|
|
|
|
|
|
|
| 4 |
|
|
|
|
| 5 |
from cleantext import clean
|
| 6 |
import gradio as gr
|
| 7 |
from tqdm.auto import tqdm
|
| 8 |
from transformers import pipeline
|
| 9 |
from transformers import AutoModelForSequenceClassification, AutoTokenizer
|
| 10 |
|
|
|
|
|
|
|
| 11 |
|
| 12 |
checker_model_name = "textattack/roberta-base-CoLA"
|
| 13 |
corrector_model_name = "pszemraj/flan-t5-large-grammar-synthesis"
|
| 14 |
|
| 15 |
# pipelines
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
if os.environ.get("HF_DEMO_NO_USE_ONNX") is None:
|
| 19 |
-
from optimum.bettertransformer import BetterTransformer
|
| 20 |
-
|
| 21 |
-
model_hf = AutoModelForSequenceClassification.from_pretrained(checker_model_name)
|
| 22 |
-
tokenizer = AutoTokenizer.from_pretrained(checker_model_name)
|
| 23 |
-
model = BetterTransformer.transform(model_hf, keep_original_model=False)
|
| 24 |
-
|
| 25 |
-
checker = pipeline(
|
| 26 |
-
"text-classification",
|
| 27 |
-
model=model,
|
| 28 |
-
tokenizer=tokenizer,
|
| 29 |
-
)
|
| 30 |
-
else:
|
| 31 |
-
checker = pipeline(
|
| 32 |
"text-classification",
|
| 33 |
checker_model_name,
|
| 34 |
)
|
|
|
|
|
|
|
| 35 |
gc.collect()
|
|
|
|
| 36 |
if os.environ.get("HF_DEMO_NO_USE_ONNX") is None:
|
| 37 |
# load onnx runtime unless HF_DEMO_NO_USE_ONNX is set
|
| 38 |
from optimum.pipelines import pipeline
|
|
@@ -130,4 +122,4 @@ with gr.Blocks() as demo:
|
|
| 130 |
"- see the [model card](https://huggingface.co/pszemraj/flan-t5-large-grammar-synthesis) for more info"
|
| 131 |
)
|
| 132 |
gr.Markdown("- if experiencing long wait times, feel free to duplicate the space!")
|
| 133 |
-
demo.launch()
|
|
|
|
|
|
|
|
|
|
| 1 |
import gc
|
| 2 |
+
import logging
|
| 3 |
+
import os
|
| 4 |
+
import re
|
| 5 |
|
| 6 |
+
import torch
|
| 7 |
from cleantext import clean
|
| 8 |
import gradio as gr
|
| 9 |
from tqdm.auto import tqdm
|
| 10 |
from transformers import pipeline
|
| 11 |
from transformers import AutoModelForSequenceClassification, AutoTokenizer
|
| 12 |
|
| 13 |
+
logging.basicConfig(level=logging.INFO)
|
| 14 |
+
logging.info(f"torch version:\t{torch.__version__}")
|
| 15 |
|
| 16 |
checker_model_name = "textattack/roberta-base-CoLA"
|
| 17 |
corrector_model_name = "pszemraj/flan-t5-large-grammar-synthesis"
|
| 18 |
|
| 19 |
# pipelines
|
| 20 |
+
checker = pipeline(
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
"text-classification",
|
| 22 |
checker_model_name,
|
| 23 |
)
|
| 24 |
+
checker.model = torch.compile(checker.model)
|
| 25 |
+
|
| 26 |
gc.collect()
|
| 27 |
+
|
| 28 |
if os.environ.get("HF_DEMO_NO_USE_ONNX") is None:
|
| 29 |
# load onnx runtime unless HF_DEMO_NO_USE_ONNX is set
|
| 30 |
from optimum.pipelines import pipeline
|
|
|
|
| 122 |
"- see the [model card](https://huggingface.co/pszemraj/flan-t5-large-grammar-synthesis) for more info"
|
| 123 |
)
|
| 124 |
gr.Markdown("- if experiencing long wait times, feel free to duplicate the space!")
|
| 125 |
+
demo.launch(debug=True)
|