v2p1
Browse filesBeta. Revert if trash.
app.py
CHANGED
@@ -4,7 +4,8 @@ import torch
|
|
4 |
from transformers import AutoTokenizer, AutoModel
|
5 |
import plotly.graph_objects as go
|
6 |
|
7 |
-
|
|
|
8 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
9 |
model = None
|
10 |
|
@@ -16,7 +17,7 @@ if tokenizer.pad_token is None:
|
|
16 |
def get_embedding(text):
|
17 |
global model
|
18 |
if model is None:
|
19 |
-
model = AutoModel.from_pretrained(model_name).cuda()
|
20 |
model.resize_token_embeddings(len(tokenizer))
|
21 |
|
22 |
inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True, max_length=512).to('cuda')
|
@@ -47,7 +48,7 @@ def generate_text_boxes(n):
|
|
47 |
|
48 |
with gr.Blocks() as iface:
|
49 |
gr.Markdown("# 3D Embedding Comparison")
|
50 |
-
gr.Markdown("Compare the embeddings of multiple strings visualized in 3D space using
|
51 |
|
52 |
num_texts = gr.Slider(minimum=2, maximum=10, step=1, value=2, label="Number of texts to compare")
|
53 |
|
@@ -73,4 +74,4 @@ with gr.Blocks() as iface:
|
|
73 |
outputs=output
|
74 |
)
|
75 |
|
76 |
-
iface.launch()
|
|
|
4 |
from transformers import AutoTokenizer, AutoModel
|
5 |
import plotly.graph_objects as go
|
6 |
|
7 |
+
# Update the model name to Llama 3.1
|
8 |
+
model_name = "meta-llama/Meta-Llama-3.1-405B-Instruct-FP8"
|
9 |
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
10 |
model = None
|
11 |
|
|
|
17 |
def get_embedding(text):
|
18 |
global model
|
19 |
if model is None:
|
20 |
+
model = AutoModel.from_pretrained(model_name, torch_dtype=torch.float16).cuda()
|
21 |
model.resize_token_embeddings(len(tokenizer))
|
22 |
|
23 |
inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True, max_length=512).to('cuda')
|
|
|
48 |
|
49 |
with gr.Blocks() as iface:
|
50 |
gr.Markdown("# 3D Embedding Comparison")
|
51 |
+
gr.Markdown("Compare the embeddings of multiple strings visualized in 3D space using Llama 3.1.")
|
52 |
|
53 |
num_texts = gr.Slider(minimum=2, maximum=10, step=1, value=2, label="Number of texts to compare")
|
54 |
|
|
|
74 |
outputs=output
|
75 |
)
|
76 |
|
77 |
+
iface.launch()
|