valeriylo commited on
Commit
7a4feb4
1 Parent(s): f49f1a5

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +8 -6
app.py CHANGED
@@ -6,6 +6,7 @@ from langchain.document_loaders import (
6
  CSVLoader,
7
  EverNoteLoader,
8
  PDFMinerLoader,
 
9
  TextLoader,
10
  UnstructuredEmailLoader,
11
  UnstructuredEPubLoader,
@@ -44,22 +45,23 @@ LOADER_MAPPING = {
44
  ".html": (UnstructuredHTMLLoader, {}),
45
  ".md": (UnstructuredMarkdownLoader, {}),
46
  ".odt": (UnstructuredODTLoader, {}),
47
- ".pdf": (PDFMinerLoader, {}),
 
48
  ".ppt": (UnstructuredPowerPointLoader, {}),
49
  ".pptx": (UnstructuredPowerPointLoader, {}),
50
  ".txt": (TextLoader, {"encoding": "utf8"}),
51
  }
52
 
53
 
54
- #repo_name = "IlyaGusev/saiga2_13b_ggml"
55
- #model_name = "ggml-model-q4_1.bin"
56
- #embedder_name = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
57
-
58
 
 
59
  repo_name = "IlyaGusev/saiga_13b_lora_llamacpp"
60
  model_name = "ggml-model-q4_1.bin"
61
  embedder_name = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
62
-
63
  snapshot_download(repo_id=repo_name, local_dir=".", allow_patterns=model_name)
64
 
65
  model = Llama(
 
6
  CSVLoader,
7
  EverNoteLoader,
8
  PDFMinerLoader,
9
+ UnstructuredPDFLoader
10
  TextLoader,
11
  UnstructuredEmailLoader,
12
  UnstructuredEPubLoader,
 
45
  ".html": (UnstructuredHTMLLoader, {}),
46
  ".md": (UnstructuredMarkdownLoader, {}),
47
  ".odt": (UnstructuredODTLoader, {}),
48
+ #".pdf": (PDFMinerLoader, {}),
49
+ ".pdf": (UnstructuredPDFLoader, {})
50
  ".ppt": (UnstructuredPowerPointLoader, {}),
51
  ".pptx": (UnstructuredPowerPointLoader, {}),
52
  ".txt": (TextLoader, {"encoding": "utf8"}),
53
  }
54
 
55
 
56
+ repo_name = "IlyaGusev/saiga2_13b_ggml"
57
+ model_name = "ggml-model-q4_1.bin"
58
+ embedder_name = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
 
59
 
60
+ """
61
  repo_name = "IlyaGusev/saiga_13b_lora_llamacpp"
62
  model_name = "ggml-model-q4_1.bin"
63
  embedder_name = "sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
64
+ """
65
  snapshot_download(repo_id=repo_name, local_dir=".", allow_patterns=model_name)
66
 
67
  model = Llama(