Merge pull request #2922 from Honkware/main

Load Salesforce Xgen Models

docs/LLaMA-model.md

@@ -12,13 +12,7 @@ This guide will cover usage through the official `transformers` implementation.
 * Torrent: https://github.com/oobabooga/text-generation-webui/pull/530#issuecomment-1484235789
 * Direct download: https://huggingface.co/Neko-Institute-of-Science
 
-⚠️ The tokenizers for the Torrent source above and also for many LLaMA fine-tunes available on Hugging Face may be outdated, so I recommend downloading the following universal LLaMA tokenizer: 
+⚠️ The tokenizers for the Torrent source above and also for many LLaMA fine-tunes available on Hugging Face may be outdated, in particular the files called `tokenizer_config.json` and `special_tokens_map.json`. Here you can find those files: https://huggingface.co/oobabooga/llama-tokenizer
-
-```
-python download-model.py oobabooga/llama-tokenizer
-```
-
-Once downloaded, it will be automatically applied to **every** `LlamaForCausalLM` model that you try to load.
 
 ### Option 2: convert the weights yourself
 

models/config.yaml

@@ -240,3 +240,6 @@ TheBloke_WizardLM-30B-GPTQ:
   truncation_length: 8192
 .*superhot-8k:
   truncation_length: 8192
+.*xgen.*-inst:
+  truncation_length: 8192
+  instruction_template: 'Vicuna-v0'

modules/models.py

@@ -3,6 +3,7 @@ import os
 import re
 import time
 from pathlib import Path
+import hashlib
 
 import torch
 import transformers
@@ -14,7 +15,6 @@ from transformers import (
     AutoModelForSeq2SeqLM,
     AutoTokenizer,
     BitsAndBytesConfig,
-    LlamaTokenizer
 )
 
 import modules.shared as shared
@@ -91,30 +91,31 @@ def load_model(model_name, loader=None):
 
 def load_tokenizer(model_name, model):
     tokenizer = None
+    path_to_model = Path(f"{shared.args.model_dir}/{model_name}/")
     if any(s in model_name.lower() for s in ['gpt-4chan', 'gpt4chan']) and Path(f"{shared.args.model_dir}/gpt-j-6B/").exists():
         tokenizer = AutoTokenizer.from_pretrained(Path(f"{shared.args.model_dir}/gpt-j-6B/"))
-    elif model.__class__.__name__ in ['LlamaForCausalLM', 'LlamaGPTQForCausalLM', 'ExllamaHF']:
+    elif path_to_model.exists():
-        # Try to load an universal LLaMA tokenizer
+        tokenizer = AutoTokenizer.from_pretrained(
-        if not any(s in shared.model_name.lower() for s in ['llava', 'oasst']):
+            path_to_model,
-            for p in [Path(f"{shared.args.model_dir}/llama-tokenizer/"), Path(f"{shared.args.model_dir}/oobabooga_llama-tokenizer/")]:
+            trust_remote_code=shared.args.trust_remote_code,
-                if p.exists():
+            use_fast=False
-                    logger.info(f"Loading the universal LLaMA tokenizer from {p}...")
+        )
-                    tokenizer = LlamaTokenizer.from_pretrained(p, clean_up_tokenization_spaces=True)
-                    return tokenizer
 
-        # Otherwise, load it from the model folder and hope that these
+    if tokenizer.__class__.__name__ == 'LlamaTokenizer':
-        # are not outdated tokenizer files.
+        pairs = [
-        tokenizer = LlamaTokenizer.from_pretrained(Path(f"{shared.args.model_dir}/{model_name}/"), clean_up_tokenization_spaces=True)
+            ['tokenizer_config.json', '516c6167c884793a738c440e29ccb80c15e1493ffc965affc69a1a8ddef4572a'],
-        try:
+            ['special_tokens_map.json', 'ff3b4a612c4e447acb02d40071bddd989fe0da87eb5b7fe0dbadfc4f74de7531']
-            tokenizer.eos_token_id = 2
+        ]
-            tokenizer.bos_token_id = 1
+
-            tokenizer.pad_token_id = 0
+        for pair in pairs:
-        except:
+            p = path_to_model / pair[0]
-            pass
+            if p.exists():
-    else:
+                with open(p, "rb") as f:
-        path_to_model = Path(f"{shared.args.model_dir}/{model_name}/")
+                    bytes = f.read()
-        if path_to_model.exists():
+
-            tokenizer = AutoTokenizer.from_pretrained(path_to_model, trust_remote_code=shared.args.trust_remote_code)
+                file_hash = hashlib.sha256(bytes).hexdigest()
+                if file_hash != pair[1]:
+                    logger.warning(f"{p} is different from the original LlamaTokenizer file. It is either customized or outdated.")
 
     return tokenizer