Rausda6 commited on
Commit
17dc694
·
verified ·
1 Parent(s): 3b3e089

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +6 -1
app.py CHANGED
@@ -13,6 +13,10 @@ from typing import List, Dict, Optional
13
  from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig
14
  import PyPDF2
15
  import traceback
 
 
 
 
16
 
17
  #from git import Repo
18
 
@@ -316,7 +320,8 @@ Requirements:
316
  - DO NOT copy the example below , only use it as conversation reference
317
  - The podcast should be professional, in-depth, lively, witty and engaging, and hook the listener from the start.
318
  - The input text might be disorganized or unformatted. Ignore any formatting inconsistencies or irrelevant details; your task is to distill the essential points, identify key definitions, and highlight intriguing facts that would be suitable for discussion in a podcast.
319
- - The two-person conversation must be in following format, always leading with the two speaker names: Speaker 1 and Speaker 2
 
320
  Speaker 1: Hello to our podcast
321
  Speaker 2: Thanks to be here
322
  Speaker 1: ...
 
13
  from transformers import AutoTokenizer, AutoModelForCausalLM, GenerationConfig
14
  import PyPDF2
15
  import traceback
16
+ import os
17
+
18
+ # Enable persistent caching on Hugging Face Spaces (if persistent storage is enabled)
19
+ os.environ["TRANSFORMERS_CACHE"] = "/data/models"
20
 
21
  #from git import Repo
22
 
 
320
  - DO NOT copy the example below , only use it as conversation reference
321
  - The podcast should be professional, in-depth, lively, witty and engaging, and hook the listener from the start.
322
  - The input text might be disorganized or unformatted. Ignore any formatting inconsistencies or irrelevant details; your task is to distill the essential points, identify key definitions, and highlight intriguing facts that would be suitable for discussion in a podcast.
323
+ - The two-person conversation must be in following format: Each Line always leading with the two speaker names: Speaker 1 or Speaker 2. No other names allowed!
324
+ Example:
325
  Speaker 1: Hello to our podcast
326
  Speaker 2: Thanks to be here
327
  Speaker 1: ...