Spaces:
Running
on
Zero
Running
on
Zero
adding patch counts and cleaning up
Browse files
app.py
CHANGED
@@ -3,15 +3,27 @@ import gradio as gr
|
|
3 |
import torch
|
4 |
import itertools # For color cycling
|
5 |
import tiktoken # For GPT-4 tokenizer
|
6 |
-
from transformers import AutoTokenizer
|
7 |
|
8 |
# Bytelatent imports (assuming they are in the python path)
|
9 |
-
|
10 |
-
from bytelatent.
|
11 |
-
from bytelatent.
|
12 |
-
from bytelatent.
|
13 |
-
from bytelatent.
|
14 |
-
from
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
15 |
|
16 |
# --- Global Setup ---
|
17 |
|
@@ -27,113 +39,117 @@ LLAMA3_MODEL_NAME = "meta-llama/Meta-Llama-3-8B" # Or choose another variant lik
|
|
27 |
|
28 |
def create_bytelatent_highlight_data(tokenizer, patch_lengths_tensor, tokens_tensor, colors):
|
29 |
"""Generates data for gr.HighlightedText based on bytelatent patches."""
|
30 |
-
|
|
|
31 |
if patch_lengths_tensor is None or tokens_tensor is None or patch_lengths_tensor.numel() == 0:
|
32 |
return None
|
33 |
patch_lengths = patch_lengths_tensor.tolist()
|
34 |
all_tokens = tokens_tensor.tolist()
|
35 |
highlighted_data = []
|
36 |
current_token_index = 0
|
37 |
-
|
|
|
38 |
for i, length in enumerate(patch_lengths):
|
39 |
if length <= 0: continue
|
40 |
patch_token_ids = all_tokens[current_token_index : current_token_index + length]
|
41 |
if not patch_token_ids: continue
|
42 |
try: patch_text = tokenizer.decode(patch_token_ids)
|
43 |
except Exception as decode_err:
|
44 |
-
|
45 |
-
|
46 |
patch_label = f"BL Patch {i+1}"
|
47 |
highlighted_data.append((patch_text, patch_label))
|
|
|
48 |
current_token_index += length
|
|
|
|
|
49 |
if current_token_index != len(all_tokens):
|
50 |
print(f"Warning: Bytelatent token mismatch. Consumed {current_token_index}, total {len(all_tokens)}")
|
51 |
remaining_tokens = all_tokens[current_token_index:]
|
52 |
if remaining_tokens:
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
|
|
|
|
57 |
|
58 |
|
59 |
def create_tiktoken_highlight_data(prompt, colors):
|
60 |
"""Generates data for gr.HighlightedText based on tiktoken (gpt-4) tokens."""
|
61 |
-
# (Keep the function from the previous version - no changes needed)
|
62 |
try:
|
63 |
enc = tiktoken.get_encoding("cl100k_base")
|
64 |
tiktoken_ids = enc.encode(prompt)
|
65 |
highlighted_data = []
|
66 |
-
color_cycler = itertools.cycle(colors)
|
67 |
for i, token_id in enumerate(tiktoken_ids):
|
68 |
try: token_text = enc.decode([token_id])
|
69 |
except UnicodeDecodeError:
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
except Exception as e:
|
75 |
-
|
76 |
-
|
77 |
token_label = f"GPT4 Tk {i+1}"
|
78 |
highlighted_data.append((token_text, token_label))
|
79 |
-
|
80 |
-
|
|
|
81 |
except ImportError:
|
82 |
-
|
83 |
-
|
84 |
except Exception as tiktoken_err:
|
85 |
print(f"Error during tiktoken processing: {tiktoken_err}")
|
86 |
-
return [(f"Error processing with tiktoken: {str(tiktoken_err)}", "Error")]
|
87 |
|
88 |
|
89 |
def create_llama3_highlight_data(prompt, colors, model_name=LLAMA3_MODEL_NAME):
|
90 |
"""Generates data for gr.HighlightedText based on Llama 3 tokenizer."""
|
91 |
try:
|
92 |
# Load Llama 3 tokenizer from Hugging Face Hub
|
93 |
-
# This might download the tokenizer files on the first run
|
94 |
-
# May require `huggingface-cli login` if model is private or gated
|
95 |
print(f"Loading Llama 3 tokenizer: {model_name}")
|
96 |
-
|
|
|
97 |
print("Llama 3 tokenizer loaded.")
|
98 |
|
99 |
# Encode the prompt
|
100 |
llama_token_ids = tokenizer.encode(prompt)
|
101 |
|
102 |
highlighted_data = []
|
103 |
-
color_cycler = itertools.cycle(colors)
|
104 |
|
105 |
for i, token_id in enumerate(llama_token_ids):
|
106 |
try:
|
107 |
-
# Decode individual token.
|
108 |
token_text = tokenizer.decode([token_id])
|
109 |
-
# Special case: Handle potential leading space added by sentencepiece during decode
|
110 |
-
# if token_text.startswith(' '): # Check if this improves visualization
|
111 |
-
# token_text = token_text[1:] # Remove leading space visual artifact? Test this.
|
112 |
except Exception as e:
|
113 |
-
|
114 |
-
|
115 |
|
116 |
token_label = f"Llama3 Tk {i+1}" # Clearer label prefix
|
117 |
highlighted_data.append((token_text, token_label))
|
118 |
|
119 |
-
|
120 |
-
|
|
|
121 |
|
122 |
except ImportError:
|
123 |
-
|
124 |
-
|
125 |
except OSError as e:
|
126 |
# Handle errors like model not found, network issues, authentication needed
|
127 |
print(f"Error loading Llama 3 tokenizer '{model_name}': {e}")
|
|
|
128 |
if "authentication" in str(e).lower():
|
129 |
-
|
130 |
-
|
131 |
-
return [(f"Could not load Llama 3 tokenizer '{model_name}'. Check model name and network. Error: {e}", "Error")]
|
132 |
except Exception as llama_err:
|
133 |
print(f"Error during Llama 3 processing: {llama_err}")
|
134 |
import traceback
|
135 |
traceback.print_exc() # Print full traceback for debugging
|
136 |
-
return [(f"Error processing with Llama 3: {str(llama_err)}", "Error")]
|
137 |
|
138 |
|
139 |
# --- Main Processing Function ---
|
@@ -141,7 +157,7 @@ def create_llama3_highlight_data(prompt, colors, model_name=LLAMA3_MODEL_NAME):
|
|
141 |
def process_text(prompt: str, model_name: str = "blt-1b"):
|
142 |
"""
|
143 |
Processes the input prompt using ByteLatent, Tiktoken, and Llama 3,
|
144 |
-
returning visualizations and status.
|
145 |
|
146 |
Args:
|
147 |
prompt: The input text string from the Gradio interface.
|
@@ -151,100 +167,136 @@ def process_text(prompt: str, model_name: str = "blt-1b"):
|
|
151 |
A tuple containing:
|
152 |
- Matplotlib Figure for the entropy plot (or None).
|
153 |
- List of tuples for bytelatent gr.HighlightedText (or None).
|
|
|
154 |
- List of tuples for tiktoken gr.HighlightedText (or None).
|
|
|
155 |
- List of tuples for Llama 3 gr.HighlightedText (or None).
|
|
|
156 |
- Status/Error message string.
|
157 |
"""
|
158 |
fig = None
|
159 |
bl_highlighted_data = None
|
160 |
tk_highlighted_data = None
|
161 |
llama_highlighted_data = None
|
|
|
|
|
|
|
162 |
status_message = "Starting processing..."
|
163 |
|
164 |
# --- 1. Tiktoken Processing (Independent) ---
|
165 |
status_message += "\nProcessing with Tiktoken (gpt-4)..."
|
166 |
-
tk_highlighted_data = create_tiktoken_highlight_data(prompt, VIZ_COLORS)
|
167 |
if tk_highlighted_data and tk_highlighted_data[0][1] == "Error":
|
168 |
-
|
|
|
169 |
else:
|
170 |
-
|
|
|
171 |
|
172 |
# --- 2. Llama 3 Processing (Independent) ---
|
173 |
status_message += "\nProcessing with Llama 3 tokenizer..."
|
174 |
-
llama_highlighted_data = create_llama3_highlight_data(prompt, VIZ_COLORS)
|
175 |
if llama_highlighted_data and llama_highlighted_data[0][1] == "Error":
|
176 |
-
|
|
|
177 |
else:
|
178 |
-
|
|
|
179 |
|
180 |
# --- 3. Bytelatent Processing ---
|
181 |
-
|
182 |
-
|
183 |
-
|
184 |
-
|
185 |
-
|
186 |
-
|
187 |
-
|
188 |
-
|
189 |
-
|
190 |
-
|
191 |
-
|
192 |
-
|
193 |
-
|
194 |
-
|
195 |
-
|
196 |
-
|
197 |
-
|
198 |
-
|
199 |
-
|
200 |
-
|
201 |
-
|
202 |
-
|
203 |
-
|
204 |
-
|
205 |
-
|
206 |
-
|
207 |
-
|
208 |
-
|
209 |
-
|
210 |
-
|
211 |
-
|
212 |
-
|
213 |
-
|
214 |
-
|
215 |
-
|
216 |
-
|
217 |
-
|
218 |
-
|
219 |
-
|
220 |
-
|
221 |
-
|
222 |
-
|
223 |
-
|
224 |
-
|
225 |
-
|
226 |
-
|
227 |
-
|
228 |
-
|
229 |
-
|
230 |
-
|
231 |
-
|
232 |
-
|
233 |
-
|
234 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
235 |
|
236 |
# Return all generated data and the final status message
|
237 |
-
return fig, bl_highlighted_data, tk_highlighted_data, llama_highlighted_data, status_message
|
238 |
|
239 |
|
240 |
# --- Gradio Interface ---
|
241 |
|
242 |
# Create color maps for HighlightedText dynamically
|
243 |
-
MAX_EXPECTED_SEGMENTS =
|
244 |
-
common_error_map = {"Error": "#FF0000"} # Red for errors
|
245 |
|
246 |
bytelatent_color_map = {f"BL Patch {i+1}": color for i, color in zip(range(MAX_EXPECTED_SEGMENTS), itertools.cycle(VIZ_COLORS))}
|
247 |
-
bytelatent_color_map["BL Remainder"] = "#
|
248 |
|
249 |
tiktoken_color_map = {f"GPT4 Tk {i+1}": color for i, color in zip(range(MAX_EXPECTED_SEGMENTS), itertools.cycle(VIZ_COLORS))}
|
250 |
tiktoken_color_map.update(common_error_map)
|
@@ -253,71 +305,118 @@ llama3_color_map = {f"Llama3 Tk {i+1}": color for i, color in zip(range(MAX_EXPE
|
|
253 |
llama3_color_map.update(common_error_map)
|
254 |
|
255 |
|
256 |
-
with gr.Blocks(theme=gr.themes.
|
257 |
-
gr.Markdown("# BLT's Entropy Patcher Visualisation")
|
258 |
gr.Markdown(
|
259 |
-
"Enter text to visualize its segmentation according to different
|
260 |
-
"1. **BLT:** Entropy plot and text
|
261 |
"2. **Tiktoken (GPT-4):** Text segmented by `cl100k_base` tokens.\n"
|
262 |
-
"3. **Llama 3:** Text segmented by the `
|
263 |
)
|
264 |
|
265 |
with gr.Row():
|
266 |
-
|
267 |
prompt_input = gr.Textbox(
|
268 |
label="Input Prompt",
|
269 |
value="Daenerys Targaryen is in Game of Thrones, a fantasy epic by George R.R. Martin.",
|
270 |
placeholder="Enter text here...",
|
271 |
-
max_length=
|
272 |
lines=5,
|
273 |
-
info="
|
274 |
)
|
275 |
submit_button = gr.Button("Generate Visualizations", variant="primary")
|
276 |
-
status_output = gr.Textbox(label="Processing Status", interactive=False, lines=
|
277 |
-
|
278 |
-
|
279 |
-
|
280 |
-
|
281 |
-
|
282 |
-
|
283 |
-
|
284 |
-
|
285 |
-
|
286 |
-
|
287 |
-
|
288 |
-
|
289 |
-
|
290 |
-
|
291 |
-
|
292 |
-
|
293 |
-
|
294 |
-
|
295 |
-
|
296 |
-
|
297 |
-
|
298 |
-
|
299 |
-
|
300 |
-
|
301 |
-
|
302 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
303 |
|
304 |
# Define the action for the button click
|
305 |
submit_button.click(
|
306 |
fn=process_text,
|
307 |
inputs=prompt_input,
|
308 |
-
# Ensure order matches the
|
309 |
outputs=[
|
310 |
-
plot_output,
|
311 |
-
highlighted_output_bl,
|
312 |
-
|
313 |
-
|
314 |
-
|
|
|
|
|
|
|
315 |
]
|
316 |
)
|
317 |
|
318 |
# --- Launch the Gradio App ---
|
319 |
if __name__ == "__main__":
|
320 |
-
print("
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
321 |
print(f"Attempting to use Llama 3 Tokenizer: {LLAMA3_MODEL_NAME}. Ensure you have access (e.g., via `huggingface-cli login` if needed).")
|
322 |
-
|
323 |
iface.launch()
|
|
|
3 |
import torch
|
4 |
import itertools # For color cycling
|
5 |
import tiktoken # For GPT-4 tokenizer
|
6 |
+
from transformers import AutoTokenizer # For Llama3 tokenizer - AutoModel usually not needed just for tokenizer
|
7 |
|
8 |
# Bytelatent imports (assuming they are in the python path)
|
9 |
+
try:
|
10 |
+
from bytelatent.data.file_util import get_fs
|
11 |
+
from bytelatent.generate_patcher import patcher_nocache
|
12 |
+
from bytelatent.tokenizers.blt_tokenizer import BltTokenizer
|
13 |
+
from bytelatent.plotting.entropy_figure_via_matplot_lib import plot_entropies
|
14 |
+
from bytelatent.args import TrainArgs
|
15 |
+
from download_blt_weights import main as ensure_present
|
16 |
+
BLT_AVAILABLE = True
|
17 |
+
except ImportError as e:
|
18 |
+
print(f"Warning: Bytelatent libraries not found. Bytelatent functionality will be disabled. Error: {e}")
|
19 |
+
BLT_AVAILABLE = False
|
20 |
+
# Define dummy classes/functions if BLT is not available to avoid NameErrors later
|
21 |
+
class BltTokenizer: pass
|
22 |
+
class TrainArgs: pass
|
23 |
+
def patcher_nocache(*args, **kwargs): return None
|
24 |
+
def plot_entropies(*args, **kwargs): return None
|
25 |
+
def ensure_present(*args, **kwargs): pass
|
26 |
+
|
27 |
|
28 |
# --- Global Setup ---
|
29 |
|
|
|
39 |
|
40 |
def create_bytelatent_highlight_data(tokenizer, patch_lengths_tensor, tokens_tensor, colors):
|
41 |
"""Generates data for gr.HighlightedText based on bytelatent patches."""
|
42 |
+
if not BLT_AVAILABLE:
|
43 |
+
return [("Bytelatent library not available.", "Error")]
|
44 |
if patch_lengths_tensor is None or tokens_tensor is None or patch_lengths_tensor.numel() == 0:
|
45 |
return None
|
46 |
patch_lengths = patch_lengths_tensor.tolist()
|
47 |
all_tokens = tokens_tensor.tolist()
|
48 |
highlighted_data = []
|
49 |
current_token_index = 0
|
50 |
+
patch_count = 0 # Initialize patch count
|
51 |
+
# color_cycler = itertools.cycle(colors) # Moved inside loop if needed per-patch
|
52 |
for i, length in enumerate(patch_lengths):
|
53 |
if length <= 0: continue
|
54 |
patch_token_ids = all_tokens[current_token_index : current_token_index + length]
|
55 |
if not patch_token_ids: continue
|
56 |
try: patch_text = tokenizer.decode(patch_token_ids)
|
57 |
except Exception as decode_err:
|
58 |
+
print(f"Warning: Bytelatent patch decoding failed: {decode_err}")
|
59 |
+
patch_text = f"[Decode Error: {len(patch_token_ids)} tokens]"
|
60 |
patch_label = f"BL Patch {i+1}"
|
61 |
highlighted_data.append((patch_text, patch_label))
|
62 |
+
patch_count += 1 # Increment count for each valid patch added
|
63 |
current_token_index += length
|
64 |
+
|
65 |
+
# Handle remainder separately, don't count it as a 'patch'
|
66 |
if current_token_index != len(all_tokens):
|
67 |
print(f"Warning: Bytelatent token mismatch. Consumed {current_token_index}, total {len(all_tokens)}")
|
68 |
remaining_tokens = all_tokens[current_token_index:]
|
69 |
if remaining_tokens:
|
70 |
+
try: remaining_text = tokenizer.decode(remaining_tokens)
|
71 |
+
except Exception: remaining_text = f"[Decode Error: {len(remaining_tokens)} remaining tokens]"
|
72 |
+
highlighted_data.append((remaining_text, "BL Remainder"))
|
73 |
+
|
74 |
+
# Return both highlighted data and the calculated patch count
|
75 |
+
return highlighted_data, patch_count
|
76 |
|
77 |
|
78 |
def create_tiktoken_highlight_data(prompt, colors):
|
79 |
"""Generates data for gr.HighlightedText based on tiktoken (gpt-4) tokens."""
|
|
|
80 |
try:
|
81 |
enc = tiktoken.get_encoding("cl100k_base")
|
82 |
tiktoken_ids = enc.encode(prompt)
|
83 |
highlighted_data = []
|
84 |
+
# color_cycler = itertools.cycle(colors) # Moved inside loop if needed per-token
|
85 |
for i, token_id in enumerate(tiktoken_ids):
|
86 |
try: token_text = enc.decode([token_id])
|
87 |
except UnicodeDecodeError:
|
88 |
+
try:
|
89 |
+
token_bytes = enc.decode_single_token_bytes(token_id)
|
90 |
+
token_text = f"[Bytes: {token_bytes.hex()}]"
|
91 |
+
except Exception: token_text = "[Decode Error]"
|
92 |
except Exception as e:
|
93 |
+
print(f"Unexpected tiktoken decode error: {e}")
|
94 |
+
token_text = "[Decode Error]"
|
95 |
token_label = f"GPT4 Tk {i+1}"
|
96 |
highlighted_data.append((token_text, token_label))
|
97 |
+
token_count = len(tiktoken_ids)
|
98 |
+
print(f"Tiktoken processing complete. Found {token_count} tokens.")
|
99 |
+
return highlighted_data, token_count
|
100 |
except ImportError:
|
101 |
+
print("Error: tiktoken library not found. Please install it: pip install tiktoken")
|
102 |
+
return [("tiktoken library not installed.", "Error")], 0
|
103 |
except Exception as tiktoken_err:
|
104 |
print(f"Error during tiktoken processing: {tiktoken_err}")
|
105 |
+
return [(f"Error processing with tiktoken: {str(tiktoken_err)}", "Error")], 0
|
106 |
|
107 |
|
108 |
def create_llama3_highlight_data(prompt, colors, model_name=LLAMA3_MODEL_NAME):
|
109 |
"""Generates data for gr.HighlightedText based on Llama 3 tokenizer."""
|
110 |
try:
|
111 |
# Load Llama 3 tokenizer from Hugging Face Hub
|
|
|
|
|
112 |
print(f"Loading Llama 3 tokenizer: {model_name}")
|
113 |
+
# Use trust_remote_code=True if required by the specific model revision
|
114 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name) #, trust_remote_code=True)
|
115 |
print("Llama 3 tokenizer loaded.")
|
116 |
|
117 |
# Encode the prompt
|
118 |
llama_token_ids = tokenizer.encode(prompt)
|
119 |
|
120 |
highlighted_data = []
|
121 |
+
# color_cycler = itertools.cycle(colors) # Moved inside loop if needed per-token
|
122 |
|
123 |
for i, token_id in enumerate(llama_token_ids):
|
124 |
try:
|
125 |
+
# Decode individual token.
|
126 |
token_text = tokenizer.decode([token_id])
|
|
|
|
|
|
|
127 |
except Exception as e:
|
128 |
+
print(f"Unexpected Llama 3 decode error for token {token_id}: {e}")
|
129 |
+
token_text = "[Decode Error]"
|
130 |
|
131 |
token_label = f"Llama3 Tk {i+1}" # Clearer label prefix
|
132 |
highlighted_data.append((token_text, token_label))
|
133 |
|
134 |
+
token_count = len(llama_token_ids)
|
135 |
+
print(f"Llama 3 processing complete. Found {token_count} tokens.")
|
136 |
+
return highlighted_data, token_count
|
137 |
|
138 |
except ImportError:
|
139 |
+
print("Error: transformers or sentencepiece library not found. Please install them: pip install transformers sentencepiece")
|
140 |
+
return [("transformers/sentencepiece library not installed.", "Error")], 0
|
141 |
except OSError as e:
|
142 |
# Handle errors like model not found, network issues, authentication needed
|
143 |
print(f"Error loading Llama 3 tokenizer '{model_name}': {e}")
|
144 |
+
error_msg = f"Could not load Llama 3 tokenizer '{model_name}'. Check model name and network."
|
145 |
if "authentication" in str(e).lower():
|
146 |
+
error_msg = f"Authentication required for Llama 3 tokenizer '{model_name}'. Use `huggingface-cli login`."
|
147 |
+
return [(f"{error_msg} Error: {e}", "Error")], 0
|
|
|
148 |
except Exception as llama_err:
|
149 |
print(f"Error during Llama 3 processing: {llama_err}")
|
150 |
import traceback
|
151 |
traceback.print_exc() # Print full traceback for debugging
|
152 |
+
return [(f"Error processing with Llama 3: {str(llama_err)}", "Error")], 0
|
153 |
|
154 |
|
155 |
# --- Main Processing Function ---
|
|
|
157 |
def process_text(prompt: str, model_name: str = "blt-1b"):
|
158 |
"""
|
159 |
Processes the input prompt using ByteLatent, Tiktoken, and Llama 3,
|
160 |
+
returning visualizations, counts, and status.
|
161 |
|
162 |
Args:
|
163 |
prompt: The input text string from the Gradio interface.
|
|
|
167 |
A tuple containing:
|
168 |
- Matplotlib Figure for the entropy plot (or None).
|
169 |
- List of tuples for bytelatent gr.HighlightedText (or None).
|
170 |
+
- Integer count of bytelatent patches.
|
171 |
- List of tuples for tiktoken gr.HighlightedText (or None).
|
172 |
+
- Integer count of tiktoken tokens.
|
173 |
- List of tuples for Llama 3 gr.HighlightedText (or None).
|
174 |
+
- Integer count of Llama 3 tokens.
|
175 |
- Status/Error message string.
|
176 |
"""
|
177 |
fig = None
|
178 |
bl_highlighted_data = None
|
179 |
tk_highlighted_data = None
|
180 |
llama_highlighted_data = None
|
181 |
+
bl_count = 0
|
182 |
+
tk_count = 0
|
183 |
+
llama_count = 0
|
184 |
status_message = "Starting processing..."
|
185 |
|
186 |
# --- 1. Tiktoken Processing (Independent) ---
|
187 |
status_message += "\nProcessing with Tiktoken (gpt-4)..."
|
188 |
+
tk_highlighted_data, tk_count_calc = create_tiktoken_highlight_data(prompt, VIZ_COLORS)
|
189 |
if tk_highlighted_data and tk_highlighted_data[0][1] == "Error":
|
190 |
+
status_message += f"\nTiktoken Error: {tk_highlighted_data[0][0]}"
|
191 |
+
tk_count = 0 # Ensure count is 0 on error
|
192 |
else:
|
193 |
+
tk_count = tk_count_calc # Assign calculated count
|
194 |
+
status_message += f"\nTiktoken processing successful ({tk_count} tokens)."
|
195 |
|
196 |
# --- 2. Llama 3 Processing (Independent) ---
|
197 |
status_message += "\nProcessing with Llama 3 tokenizer..."
|
198 |
+
llama_highlighted_data, llama_count_calc = create_llama3_highlight_data(prompt, VIZ_COLORS)
|
199 |
if llama_highlighted_data and llama_highlighted_data[0][1] == "Error":
|
200 |
+
status_message += f"\nLlama 3 Error: {llama_highlighted_data[0][0]}"
|
201 |
+
llama_count = 0 # Ensure count is 0 on error
|
202 |
else:
|
203 |
+
llama_count = llama_count_calc # Assign calculated count
|
204 |
+
status_message += f"\nLlama 3 processing successful ({llama_count} tokens)."
|
205 |
|
206 |
# --- 3. Bytelatent Processing ---
|
207 |
+
if BLT_AVAILABLE:
|
208 |
+
try:
|
209 |
+
status_message += f"\nLoading Bytelatent entropy model for '{model_name}'..."
|
210 |
+
# (Bytelatent loading code remains the same)
|
211 |
+
consolidated_path = os.path.join("hf-weights", model_name)
|
212 |
+
train_args_path = os.path.join(consolidated_path, "params.json")
|
213 |
+
if not os.path.exists(train_args_path): raise FileNotFoundError(f"BLT training args not found at {train_args_path}.")
|
214 |
+
fs = get_fs(train_args_path); train_args = TrainArgs.model_validate_json(fs.read_text(train_args_path))
|
215 |
+
bl_tokenizer = train_args.data.tokenizer_args.build(); assert isinstance(bl_tokenizer, BltTokenizer)
|
216 |
+
patcher_args = train_args.data.patcher_args.model_copy(deep=True); patcher_args.realtime_patching = True
|
217 |
+
device = "cuda" if torch.cuda.is_available() else "cpu"; print(f"Using BLT device: {device}")
|
218 |
+
patcher_args.patching_device = device; patcher_args.device = device
|
219 |
+
entropy_model_dir = os.path.join(consolidated_path, "entropy_model")
|
220 |
+
if not os.path.exists(entropy_model_dir): raise FileNotFoundError(f"Entropy model directory not found at {entropy_model_dir}.")
|
221 |
+
patcher_args.entropy_model_checkpoint_dir = entropy_model_dir; bl_patcher = patcher_args.build()
|
222 |
+
status_message += "\nBytelatent entropy model loaded."
|
223 |
+
|
224 |
+
# --- Processing ---
|
225 |
+
status_message += "\nRunning Bytelatent entropy model patching..."
|
226 |
+
print(f"Processing prompt with entropy model: '{prompt}'")
|
227 |
+
prompt_bytes = prompt.encode('utf-8')
|
228 |
+
max_bytes = 512 # Define max bytes
|
229 |
+
if len(prompt_bytes) > max_bytes:
|
230 |
+
print(f"Warning: Prompt exceeds {max_bytes} bytes ({len(prompt_bytes)}). Truncating for entropy model.")
|
231 |
+
# Find the byte position that corresponds to the last full character within the limit
|
232 |
+
# This avoids splitting a multi-byte character
|
233 |
+
try:
|
234 |
+
last_char_pos = prompt_bytes[:max_bytes].rfind(b' ') # Simple whitespace split point find, might not be perfect
|
235 |
+
if last_char_pos == -1: # If no space, truncate hard (less ideal)
|
236 |
+
prompt_bl = prompt_bytes[:max_bytes].decode('utf-8', errors='ignore')
|
237 |
+
else:
|
238 |
+
prompt_bl = prompt_bytes[:last_char_pos].decode('utf-8', errors='ignore')
|
239 |
+
|
240 |
+
except Exception: # Fallback to simple truncation on decode errors
|
241 |
+
prompt_bl = prompt_bytes[:max_bytes].decode('utf-8', errors='ignore')
|
242 |
+
|
243 |
+
status_message += f"\nWarning: Prompt truncated to approx {len(prompt_bl.encode('utf-8'))} bytes for Bytelatent entropy model."
|
244 |
+
else:
|
245 |
+
prompt_bl = prompt
|
246 |
+
|
247 |
+
results = patcher_nocache([prompt_bl], tokenizer=bl_tokenizer, patcher=bl_patcher)
|
248 |
+
|
249 |
+
if not results:
|
250 |
+
print("Bytelatent entropy processing returned no results.")
|
251 |
+
status_message += "\nBytelatent entropy model warning: Processing completed, but no results were generated."
|
252 |
+
bl_highlighted_data = [("No patches generated.", "Info")]
|
253 |
+
bl_count = 0
|
254 |
+
else:
|
255 |
+
batch_patch_lengths, batch_scores, batch_tokens = results
|
256 |
+
patch_lengths, scores, tokens = batch_patch_lengths[0], batch_scores[0], batch_tokens[0]
|
257 |
+
# --- Visualization Data Generation ---
|
258 |
+
try: decoded_output_for_plot = bl_tokenizer.decode(tokens.tolist())
|
259 |
+
except Exception as decode_err:
|
260 |
+
print(f"Warning: Error decoding full sequence for plot: {decode_err}")
|
261 |
+
decoded_output_for_plot = prompt_bl # Use truncated prompt for plot if decode fails
|
262 |
+
|
263 |
+
fig = plot_entropies(patch_lengths, scores, decoded_output_for_plot, threshold=bl_patcher.threshold)
|
264 |
+
bl_highlighted_data, bl_count_calc = create_bytelatent_highlight_data(bl_tokenizer, patch_lengths, tokens, VIZ_COLORS)
|
265 |
+
bl_count = bl_count_calc # Assign calculated count
|
266 |
+
|
267 |
+
status_message += f"\nBytelatent entropy model processing and visualization successful ({bl_count} patches)."
|
268 |
+
print("Bytelatent Entropy model processing and decoding complete.")
|
269 |
+
|
270 |
+
except FileNotFoundError as e:
|
271 |
+
print(f"Bytelatent Error: {e}")
|
272 |
+
status_message += f"\nBytelatent FileNotFoundError: {str(e)}"
|
273 |
+
bl_highlighted_data = [(f"Bytelatent Error: {e}", "Error")]
|
274 |
+
bl_count = 0
|
275 |
+
except Exception as e:
|
276 |
+
print(f"An unexpected Bytelatent error occurred: {e}")
|
277 |
+
import traceback
|
278 |
+
traceback.print_exc()
|
279 |
+
status_message += f"\nBytelatent Unexpected Error: {str(e)}"
|
280 |
+
bl_highlighted_data = [(f"Bytelatent Error: {e}", "Error")]
|
281 |
+
bl_count = 0
|
282 |
+
else:
|
283 |
+
status_message += "\nBytelatent processing skipped (library not found)."
|
284 |
+
bl_highlighted_data = [("Bytelatent library not available.", "Error")]
|
285 |
+
bl_count = 0
|
286 |
+
fig = None # Ensure fig is None if BLT is skipped
|
287 |
|
288 |
# Return all generated data and the final status message
|
289 |
+
return fig, bl_highlighted_data, bl_count, tk_highlighted_data, tk_count, llama_highlighted_data, llama_count, status_message
|
290 |
|
291 |
|
292 |
# --- Gradio Interface ---
|
293 |
|
294 |
# Create color maps for HighlightedText dynamically
|
295 |
+
MAX_EXPECTED_SEGMENTS = 2000 # Increased max segments further just in case
|
296 |
+
common_error_map = {"Error": "#FF0000", "Info": "#808080"} # Red for errors, Gray for info
|
297 |
|
298 |
bytelatent_color_map = {f"BL Patch {i+1}": color for i, color in zip(range(MAX_EXPECTED_SEGMENTS), itertools.cycle(VIZ_COLORS))}
|
299 |
+
bytelatent_color_map["BL Remainder"] = "#AAAAAA"; bytelatent_color_map.update(common_error_map)
|
300 |
|
301 |
tiktoken_color_map = {f"GPT4 Tk {i+1}": color for i, color in zip(range(MAX_EXPECTED_SEGMENTS), itertools.cycle(VIZ_COLORS))}
|
302 |
tiktoken_color_map.update(common_error_map)
|
|
|
305 |
llama3_color_map.update(common_error_map)
|
306 |
|
307 |
|
308 |
+
with gr.Blocks(theme=gr.themes.Origin()) as iface:
|
309 |
+
gr.Markdown("# BLT's Entropy-based Patcher vs. Tokenizer Visualisation")
|
310 |
gr.Markdown(
|
311 |
+
"Enter text to visualize its segmentation according to different methods:\n"
|
312 |
+
"1. **Byte Latent Transformer (BLT):** Entropy-based patching plot and patched text (_for this space ONLY_ - limited to ~512 bytes).\n"
|
313 |
"2. **Tiktoken (GPT-4):** Text segmented by `cl100k_base` tokens.\n"
|
314 |
+
f"3. **Llama 3:** Text segmented by the `{LLAMA3_MODEL_NAME}` tokenizer."
|
315 |
)
|
316 |
|
317 |
with gr.Row():
|
318 |
+
with gr.Column(scale=1): # Input Column
|
319 |
prompt_input = gr.Textbox(
|
320 |
label="Input Prompt",
|
321 |
value="Daenerys Targaryen is in Game of Thrones, a fantasy epic by George R.R. Martin.",
|
322 |
placeholder="Enter text here...",
|
323 |
+
max_length=512, # Allow even longer input, Bytelatent will truncate
|
324 |
lines=5,
|
325 |
+
info="For this space ONLY, processing is limited to ~512 bytes."
|
326 |
)
|
327 |
submit_button = gr.Button("Generate Visualizations", variant="primary")
|
328 |
+
status_output = gr.Textbox(label="Processing Status", interactive=False, lines=7) # Increased lines slightly
|
329 |
+
|
330 |
+
with gr.Column(scale=2): # Output Column
|
331 |
+
|
332 |
+
# --- Bytelatent Output Area ---
|
333 |
+
with gr.Row(equal_height=False): # Use Row to place title and count together
|
334 |
+
gr.Markdown("### BLT Entropy Patcher Output (`blt_main_entropy_100m_512w`)")
|
335 |
+
|
336 |
+
bl_count_output = gr.Number(label="Patch Count", value=0, interactive=False, scale=1, step=1) # Added Number output
|
337 |
+
highlighted_output_bl = gr.HighlightedText(
|
338 |
+
label="BLT's Entropy-based Patches",
|
339 |
+
color_map=bytelatent_color_map,
|
340 |
+
show_legend=False, # Legend can get very long
|
341 |
+
# show_label=False, # Hide the HighlightedText label as we have the markdown title
|
342 |
+
show_inline_category=False,
|
343 |
+
# container=False, # Reduces vertical space slightly
|
344 |
+
)
|
345 |
+
plot_output = gr.Plot(label="Entropy vs. Token Index", show_label=True)
|
346 |
+
|
347 |
+
# --- Tiktoken Output Area ---
|
348 |
+
with gr.Row(equal_height=False):
|
349 |
+
gr.Markdown("### Tiktoken Output (`cl100k_base`)")
|
350 |
+
|
351 |
+
tk_count_output = gr.Number(label="Token Count", value=0, interactive=False, scale=1, step=1) # Added Number output
|
352 |
+
highlighted_output_tk = gr.HighlightedText(
|
353 |
+
label="Tiktoken Segmented Text",
|
354 |
+
color_map=tiktoken_color_map,
|
355 |
+
show_legend=False,
|
356 |
+
show_inline_category=False,
|
357 |
+
# show_label=False,
|
358 |
+
# container=False,
|
359 |
+
)
|
360 |
+
|
361 |
+
# --- Llama 3 Output Area ---
|
362 |
+
with gr.Row(equal_height=False):
|
363 |
+
gr.Markdown(f"### Llama 3 Output (`{LLAMA3_MODEL_NAME}`)")
|
364 |
+
|
365 |
+
llama_count_output = gr.Number(label="Token Count", value=0, interactive=False, scale=1, step=1) # Added Number output
|
366 |
+
highlighted_output_llama = gr.HighlightedText(
|
367 |
+
label="Llama 3 Segmented Text",
|
368 |
+
color_map=llama3_color_map,
|
369 |
+
show_legend=False,
|
370 |
+
show_inline_category=False,
|
371 |
+
# show_label=False,
|
372 |
+
# container=False,
|
373 |
+
)
|
374 |
|
375 |
# Define the action for the button click
|
376 |
submit_button.click(
|
377 |
fn=process_text,
|
378 |
inputs=prompt_input,
|
379 |
+
# Ensure order matches the 8 return values of process_text
|
380 |
outputs=[
|
381 |
+
plot_output, # fig
|
382 |
+
highlighted_output_bl, # bl_highlighted_data
|
383 |
+
bl_count_output, # bl_count <-- New
|
384 |
+
highlighted_output_tk, # tk_highlighted_data
|
385 |
+
tk_count_output, # tk_count <-- New
|
386 |
+
highlighted_output_llama,# llama_highlighted_data
|
387 |
+
llama_count_output, # llama_count <-- New
|
388 |
+
status_output # status_message
|
389 |
]
|
390 |
)
|
391 |
|
392 |
# --- Launch the Gradio App ---
|
393 |
if __name__ == "__main__":
|
394 |
+
print("Checking required libraries...")
|
395 |
+
try:
|
396 |
+
import tiktoken
|
397 |
+
print("- tiktoken found.")
|
398 |
+
except ImportError:
|
399 |
+
print("WARNING: 'tiktoken' not found. GPT-4 visualization will fail. Install with: pip install tiktoken")
|
400 |
+
try:
|
401 |
+
import transformers
|
402 |
+
import sentencepiece
|
403 |
+
print("- transformers found.")
|
404 |
+
print("- sentencepiece found.")
|
405 |
+
except ImportError:
|
406 |
+
print("WARNING: 'transformers' or 'sentencepiece' not found. Llama 3 visualization will fail. Install with: pip install transformers sentencepiece")
|
407 |
+
|
408 |
+
if BLT_AVAILABLE:
|
409 |
+
print("- Bytelatent libraries found.")
|
410 |
+
# Ensure bytelatent model is present only if library is available
|
411 |
+
try:
|
412 |
+
print("Ensuring Bytelatent model 'blt-1b' weights are present...")
|
413 |
+
ensure_present(["blt-1b"])
|
414 |
+
print("Bytelatent model check complete.")
|
415 |
+
except Exception as blt_dl_err:
|
416 |
+
print(f"WARNING: Failed to ensure Bytelatent model presence: {blt_dl_err}")
|
417 |
+
else:
|
418 |
+
print("INFO: Bytelatent libraries not found, skipping related functionality.")
|
419 |
+
|
420 |
print(f"Attempting to use Llama 3 Tokenizer: {LLAMA3_MODEL_NAME}. Ensure you have access (e.g., via `huggingface-cli login` if needed).")
|
421 |
+
print("Launching Gradio interface...")
|
422 |
iface.launch()
|