Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import zlib
|
2 |
+
|
3 |
+
import gradio as gr
|
4 |
+
|
5 |
+
import gptzip
|
6 |
+
import transformers
|
7 |
+
|
8 |
+
model = "EleutherAI/pythia-410m"
|
9 |
+
|
10 |
+
lm = transformers.AutoModelForCausalLM.from_pretrained(model)
|
11 |
+
tokenizer = transformers.AutoTokenizer.from_pretrained(model)
|
12 |
+
coder = gptzip.ArithmeticCoder(lm=lm, tokenizer=tokenizer)
|
13 |
+
|
14 |
+
def compress(string: str) -> str:
|
15 |
+
code, num_padded_bits = coder.encode(
|
16 |
+
string,
|
17 |
+
return_num_padded_bits=True,
|
18 |
+
)
|
19 |
+
gzip_str = zlib.compress(string.encode())
|
20 |
+
gzip_bytes = len(gzip_str)
|
21 |
+
return f"Num tokens {len(tokenizer.encode(string))} || GZ: {gzip_bytes} bytes: {gzip_str} || AC: {len(code)} bytes: {code}"
|
22 |
+
|
23 |
+
demo = gr.Interface(fn=compress, inputs="text", outputs="text")
|
24 |
+
demo.launch()
|