tacit/nanochat
1
0
Fork 0
mirror of https://github.com/karpathy/nanochat.git synced 2025-12-06 04:12:13 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

allow the tokenizer visualize_tokenization to also print the exact token id. you can never be paranoid enough

Browse Source
This commit is contained in:
Andrej Karpathy 2025-10-24 13:27:05 +00:00
parent 5eeb2b6ef9
commit cc3636b01c
1 changed files with 4 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
nanochat/tokenizer.py
View File

@ -341,16 +341,19 @@ class RustBPETokenizer:
mask = mask[:max_tokens] mask = mask[:max_tokens]
return ids, mask return ids, mask
def visualize_tokenization(self, ids, mask): def visualize_tokenization(self, ids, mask, with_token_id=False):
"""Small helper function useful in debugging: visualize the tokenization of render_conversation""" """Small helper function useful in debugging: visualize the tokenization of render_conversation"""
RED = '\033[91m' RED = '\033[91m'
GREEN = '\033[92m' GREEN = '\033[92m'
RESET = '\033[0m' RESET = '\033[0m'
GRAY = '\033[90m'
tokens = [] tokens = []
for i, (token_id, mask_val) in enumerate(zip(ids, mask)): for i, (token_id, mask_val) in enumerate(zip(ids, mask)):
token_str = self.decode([token_id]) token_str = self.decode([token_id])
color = GREEN if mask_val == 1 else RED color = GREEN if mask_val == 1 else RED
tokens.append(f"{color}{token_str}{RESET}") tokens.append(f"{color}{token_str}{RESET}")
if with_token_id:
tokens.append(f"{GRAY}({token_id}){RESET}")
return '|'.join(tokens) return '|'.join(tokens)
def render_for_completion(self, conversation): def render_for_completion(self, conversation):