xzuyn commited on
Commit
b67636b
1 Parent(s): 92761a6

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +3 -1
app.py CHANGED
@@ -7,6 +7,7 @@ gpt_neox_tokenizer = AutoTokenizer.from_pretrained("EleutherAI/gpt-neox-20b")
7
  llama_tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/llama-tokenizer")
8
  falcon_tokenizer = AutoTokenizer.from_pretrained("tiiuae/falcon-7b")
9
  phi2_tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-2")
 
10
 
11
  def tokenize(input_text):
12
  gpt2_tokens = gpt2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
@@ -14,8 +15,9 @@ def tokenize(input_text):
14
  llama_tokens = llama_tokenizer(input_text, add_special_tokens=True)["input_ids"]
15
  falcon_tokens = falcon_tokenizer(input_text, add_special_tokens=True)["input_ids"]
16
  phi2_tokens = phi2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
 
17
 
18
- return f"GPT-2/GPT-J: {len(gpt2_tokens)}\nGPT-NeoX: {len(gpt_neox_tokens)}\nLLaMa: {len(llama_tokens)}\nFalcon: {len(falcon_tokens)}\nPhi-2: {len(phi2_tokens)}"
19
 
20
  iface = gr.Interface(fn=tokenize, inputs=gr.Textbox(lines=7), outputs="text")
21
  iface.launch()
 
7
  llama_tokenizer = AutoTokenizer.from_pretrained("hf-internal-testing/llama-tokenizer")
8
  falcon_tokenizer = AutoTokenizer.from_pretrained("tiiuae/falcon-7b")
9
  phi2_tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-2")
10
+ t5_tokenizer = AutoTokenizer.from_pretrained("google/flan-t5-xxl")
11
 
12
  def tokenize(input_text):
13
  gpt2_tokens = gpt2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
 
15
  llama_tokens = llama_tokenizer(input_text, add_special_tokens=True)["input_ids"]
16
  falcon_tokens = falcon_tokenizer(input_text, add_special_tokens=True)["input_ids"]
17
  phi2_tokens = phi2_tokenizer(input_text, add_special_tokens=True)["input_ids"]
18
+ t5_tokens = t5_tokenizer(input_text, add_special_tokens=True)["input_ids"]
19
 
20
+ return f"GPT-2/GPT-J: {len(gpt2_tokens)}\nGPT-NeoX: {len(gpt_neox_tokens)}\nLLaMa: {len(llama_tokens)}\nFalcon: {len(falcon_tokens)}\nPhi-2: {len(phi2_tokens)}\nT5: {len(t5_tokens)}"
21
 
22
  iface = gr.Interface(fn=tokenize, inputs=gr.Textbox(lines=7), outputs="text")
23
  iface.launch()