Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -2,18 +2,6 @@ from transformers import AutoTokenizer
|
|
| 2 |
import gradio as gr
|
| 3 |
|
| 4 |
|
| 5 |
-
def load_tokenizers():
|
| 6 |
-
llama1_tokenizer = AutoTokenizer.from_pretrained("huggyllama/llama-7b")
|
| 7 |
-
llama2_tokenizer = AutoTokenizer.from_pretrained("TheBloke/Llama-2-7B-fp16")
|
| 8 |
-
mistral_tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
|
| 9 |
-
gpt2_tokenizer = AutoTokenizer.from_pretrained("gpt2")
|
| 10 |
-
gpt_neox_tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neox-20b")
|
| 11 |
-
falcon_tokenizer = AutoTokenizer.from_pretrained("tiiuae/falcon-7b")
|
| 12 |
-
phi2_tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-2")
|
| 13 |
-
t5_tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-xxl")
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
def tokenize(input_text):
|
| 18 |
llama1_tokens = llama1_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
| 19 |
llama2_tokens = llama2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
|
@@ -23,13 +11,19 @@ def tokenize(input_text):
|
|
| 23 |
falcon_tokens = falcon_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
| 24 |
phi2_tokens = phi2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
| 25 |
t5_tokens = t5_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
| 26 |
-
|
| 27 |
|
| 28 |
return f"LLaMa-1: {len(llama1_tokens)}\nLLaMa-2: {len(llama2_tokens)}\nMistral: {len(mistral_tokens)}\nGPT-2/GPT-J: {len(gpt2_tokens)}\nGPT-NeoX: {len(gpt_neox_tokens)}\nFalcon: {len(falcon_tokens)}\nPhi-2: {len(phi2_tokens)}\nT5: {len(t5_tokens)}"
|
| 29 |
|
| 30 |
|
| 31 |
if __name__ == "__main__":
|
| 32 |
-
|
| 33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
iface = gr.Interface(fn=tokenize, inputs=gr.Textbox(lines=7), outputs="text")
|
| 35 |
-
iface.launch()
|
|
|
|
| 2 |
import gradio as gr
|
| 3 |
|
| 4 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
def tokenize(input_text):
|
| 6 |
llama1_tokens = llama1_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
| 7 |
llama2_tokens = llama2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
|
|
|
| 11 |
falcon_tokens = falcon_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
| 12 |
phi2_tokens = phi2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
| 13 |
t5_tokens = t5_tokenizer(input_text, add_special_tokens=True)["input_ids"]
|
|
|
|
| 14 |
|
| 15 |
return f"LLaMa-1: {len(llama1_tokens)}\nLLaMa-2: {len(llama2_tokens)}\nMistral: {len(mistral_tokens)}\nGPT-2/GPT-J: {len(gpt2_tokens)}\nGPT-NeoX: {len(gpt_neox_tokens)}\nFalcon: {len(falcon_tokens)}\nPhi-2: {len(phi2_tokens)}\nT5: {len(t5_tokens)}"
|
| 16 |
|
| 17 |
|
| 18 |
if __name__ == "__main__":
|
| 19 |
+
llama1_tokenizer = AutoTokenizer.from_pretrained("huggyllama/llama-7b")
|
| 20 |
+
llama2_tokenizer = AutoTokenizer.from_pretrained("TheBloke/Llama-2-7B-fp16")
|
| 21 |
+
mistral_tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
|
| 22 |
+
gpt2_tokenizer = AutoTokenizer.from_pretrained("gpt2")
|
| 23 |
+
gpt_neox_tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neox-20b")
|
| 24 |
+
falcon_tokenizer = AutoTokenizer.from_pretrained("tiiuae/falcon-7b")
|
| 25 |
+
phi2_tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-2")
|
| 26 |
+
t5_tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-xxl")
|
| 27 |
+
|
| 28 |
iface = gr.Interface(fn=tokenize, inputs=gr.Textbox(lines=7), outputs="text")
|
| 29 |
+
iface.launch()
|