nandigama/Hands-On-Large-Language-Models
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Fix loading Phi-3 (#48)

Browse Source
This commit is contained in:
Maarten Grootendorst
2025-04-04 08:34:35 +02:00
committed by GitHub
parent e55a0738aa
commit aabef6b341
6 changed files with 9 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
chapter12/Chapter 12 - Fine-tuning Generation Models.ipynb
View File

@@ -859,7 +859,7 @@
"model.config.pretraining_tp = 1\n",
"\n",
"# Load LLaMA tokenizer\n",
"tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)\n",
"tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=False)\n",
"tokenizer.pad_token = \"<PAD>\"\n",
"tokenizer.padding_side = \"left\""
]
@@ -1598,7 +1598,7 @@
"\n",
"# Load LLaMA tokenizer\n",
"model_name = \"TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T\"\n",
"tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)\n",
"tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=False)\n",
"tokenizer.pad_token = \"<PAD>\"\n",
"tokenizer.padding_side = \"left\""
]