tanner/min-dalle-test
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

read tokenizer files with utf8 encoding

Browse Source
This commit is contained in:
Brett Kuprel 2022-06-29 14:18:23 -04:00
parent 742d3609b0
commit 61cc99c13c
1 changed files with 8 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
min_dalle/min_dalle.py
View File

@ -13,12 +13,17 @@ class MinDalle:
model_path = os.path.join('pretrained', model_name) model_path = os.path.join('pretrained', model_name)
print("reading files from {}".format(model_path)) print("reading files from {}".format(model_path))
with open(os.path.join(model_path, 'config.json'), 'r') as f: config_path = os.path.join(model_path, 'config.json')
vocab_path = os.path.join(model_path, 'vocab.json')
merges_path = os.path.join(model_path, 'merges.txt')
with open(config_path, 'r', encoding='utf8') as f:
self.config = json.load(f) self.config = json.load(f)
with open(os.path.join(model_path, 'vocab.json'), 'r') as f: with open(vocab_path, 'r', encoding='utf8') as f:
vocab = json.load(f) vocab = json.load(f)
with open(os.path.join(model_path, 'merges.txt'), 'r') as f: with open(merges_path, 'r', encoding='utf8') as f:
merges = f.read().split("\n")[1:-1] merges = f.read().split("\n")[1:-1]
self.model_params = load_dalle_bart_flax_params(model_path) self.model_params = load_dalle_bart_flax_params(model_path)
self.tokenizer = TextTokenizer(vocab, merges) self.tokenizer = TextTokenizer(vocab, merges)