Spaces:
Running
on
Zero
Running
on
Zero
Update app.py
Browse files
app.py
CHANGED
@@ -1,4 +1,4 @@
|
|
1 |
-
# app
|
2 |
|
3 |
from bert_handler import create_handler_from_checkpoint
|
4 |
import torch
|
@@ -6,40 +6,48 @@ import gradio as gr
|
|
6 |
import re
|
7 |
from pathlib import Path
|
8 |
import spaces
|
|
|
9 |
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
14 |
-
|
15 |
-
|
|
|
|
|
|
|
16 |
|
17 |
-
|
18 |
-
|
19 |
-
|
|
|
|
|
20 |
|
21 |
-
|
|
|
22 |
with torch.no_grad():
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
|
27 |
-
|
28 |
-
|
29 |
-
|
|
|
|
|
30 |
|
31 |
-
|
32 |
-
|
33 |
-
results.append({
|
34 |
-
"Position": i,
|
35 |
-
"Masked Token": MASK_TOKEN,
|
36 |
-
"Predicted": predicted_ids[i],
|
37 |
-
"Original": original_ids[i] if i < len(original_ids) else "",
|
38 |
-
"Match": "✅" if predicted_ids[i] == original_ids[i] else "❌"
|
39 |
-
})
|
40 |
|
41 |
-
|
42 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
43 |
|
44 |
symbolic_roles = [
|
45 |
"<subject>", "<subject1>", "<subject2>", "<pose>", "<emotion>",
|
@@ -50,27 +58,23 @@ symbolic_roles = [
|
|
50 |
"<fabric>", "<jewelry>"
|
51 |
]
|
52 |
|
53 |
-
# Load from official hosted checkpoint
|
54 |
-
checkpoint_path = "./bert-beatrix-2048"
|
55 |
-
handler, model, tokenizer = create_handler_from_checkpoint(checkpoint_path)
|
56 |
-
model = model.eval().cuda()
|
57 |
-
|
58 |
def build_interface():
|
59 |
with gr.Blocks() as demo:
|
60 |
-
gr.Markdown("##
|
61 |
with gr.Row():
|
62 |
with gr.Column():
|
63 |
-
input_text = gr.Textbox(label="Symbolic
|
64 |
selected_roles = gr.CheckboxGroup(
|
65 |
choices=symbolic_roles,
|
66 |
-
label="
|
67 |
)
|
68 |
-
run_btn = gr.Button("
|
69 |
with gr.Column():
|
70 |
-
|
71 |
-
|
|
|
72 |
|
73 |
-
run_btn.click(fn=
|
74 |
|
75 |
return demo
|
76 |
|
|
|
1 |
+
# Updating the app to use only the encoder from the model, ensuring symbolic support
|
2 |
|
3 |
from bert_handler import create_handler_from_checkpoint
|
4 |
import torch
|
|
|
6 |
import re
|
7 |
from pathlib import Path
|
8 |
import spaces
|
9 |
+
from huggingface_hub import snapshot_download
|
10 |
|
11 |
+
# Load checkpoint using BERTHandler (loads tokenizer and full model)
|
12 |
+
checkpoint_path = snapshot_download(
|
13 |
+
repo_id="AbstractPhil/bert-beatrix-2048",
|
14 |
+
revision="main",
|
15 |
+
local_dir="bert-beatrix-2048",
|
16 |
+
local_dir_use_symlinks=False
|
17 |
+
)
|
18 |
+
handler, model, tokenizer = create_handler_from_checkpoint(checkpoint_path)
|
19 |
+
model = model.eval().cuda()
|
20 |
|
21 |
+
# Extract encoder only (NomicBertModel -> encoder)
|
22 |
+
encoder = model.bert.encoder
|
23 |
+
embeddings = model.bert.embeddings
|
24 |
+
emb_ln = model.bert.emb_ln
|
25 |
+
emb_drop = model.bert.emb_drop
|
26 |
|
27 |
+
@spaces.GPU
|
28 |
+
def encode_and_predict(text: str, selected_roles: list[str]):
|
29 |
with torch.no_grad():
|
30 |
+
inputs = tokenizer(text, return_tensors="pt").to("cuda")
|
31 |
+
input_ids = inputs.input_ids
|
32 |
+
attention_mask = inputs.attention_mask
|
33 |
|
34 |
+
# Run embedding + encoder pipeline
|
35 |
+
x = embeddings(input_ids)
|
36 |
+
x = emb_ln(x)
|
37 |
+
x = emb_drop(x)
|
38 |
+
encoded = encoder(x, attention_mask=attention_mask.bool())
|
39 |
|
40 |
+
symbolic_ids = [tokenizer.convert_tokens_to_ids(tok) for tok in selected_roles]
|
41 |
+
symbolic_mask = torch.isin(input_ids, torch.tensor(symbolic_ids, device=input_ids.device))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
42 |
|
43 |
+
masked_tokens = [tokenizer.convert_ids_to_tokens([tid])[0] for tid in input_ids[0] if tid in symbolic_ids]
|
44 |
+
role_reprs = encoded[symbolic_mask].mean(dim=0) if symbolic_mask.any() else torch.zeros_like(encoded[0, 0])
|
45 |
+
|
46 |
+
return {
|
47 |
+
"Symbolic Tokens": masked_tokens,
|
48 |
+
"Embedding Norm": f"{role_reprs.norm().item():.4f}",
|
49 |
+
"Symbolic Token Count": symbolic_mask.sum().item(),
|
50 |
+
}
|
51 |
|
52 |
symbolic_roles = [
|
53 |
"<subject>", "<subject1>", "<subject2>", "<pose>", "<emotion>",
|
|
|
58 |
"<fabric>", "<jewelry>"
|
59 |
]
|
60 |
|
|
|
|
|
|
|
|
|
|
|
61 |
def build_interface():
|
62 |
with gr.Blocks() as demo:
|
63 |
+
gr.Markdown("## 🧠 Symbolic Encoder Inspector")
|
64 |
with gr.Row():
|
65 |
with gr.Column():
|
66 |
+
input_text = gr.Textbox(label="Input with Symbolic Tokens", lines=3)
|
67 |
selected_roles = gr.CheckboxGroup(
|
68 |
choices=symbolic_roles,
|
69 |
+
label="Which symbolic tokens should be traced?"
|
70 |
)
|
71 |
+
run_btn = gr.Button("Encode & Trace")
|
72 |
with gr.Column():
|
73 |
+
symbolic_tokens = gr.Textbox(label="Symbolic Tokens Found")
|
74 |
+
embedding_norm = gr.Textbox(label="Mean Norm of Symbolic Embeddings")
|
75 |
+
token_count = gr.Textbox(label="Count of Symbolic Tokens")
|
76 |
|
77 |
+
run_btn.click(fn=encode_and_predict, inputs=[input_text, selected_roles], outputs=[symbolic_tokens, embedding_norm, token_count])
|
78 |
|
79 |
return demo
|
80 |
|