Kezdőlap Felfedezés Súgó
Bejelentkezés
radu
/
LLamaRecipes
tükrözi: https://github.com/facebookresearch/llama-recipes.git
Figyelés 1
Kedvenc 0
Másolás 0
Fájlok Problémák 0 Wiki
Fa: d6b2f42613
Branch-ok Tag-ek
LLamaRecipes
/
recipes
/
quickstart
/
NotebookLlama
Sanyam Bhutani d6b2f42613 rename 7 hónapja
..
1B-chat-start.py d05c3ec968 chat-starter-pack 7 hónapja
1B-debating-script.py e979f0611c Create 1B-debating-script.py 7 hónapja
Bark-Testing.ipynb 0bd41ea488 Sweeps added 7 hónapja
Parler-Testing.ipynb 867d99e625 Create Parler-Testing.ipynb 7 hónapja
Prompt_testing.md eba932ba1f Update Prompt_testing.md 7 hónapja
README.md 083b59374f Add notes 7 hónapja
Step-1 PDF Pre-Processing Logic.ipynb 75b5b39c2a Happy push 7 hónapja
Step-2-70B-Rewriter.ipynb e002c0d2f2 Fix conflicts 7 hónapja
Step-2-8B-Rewriter.ipynb e002c0d2f2 Fix conflicts 7 hónapja
Step-5-TTS-Workflow.ipynb d6b2f42613 rename 7 hónapja
clean_extracted_text.txt e002c0d2f2 Fix conflicts 7 hónapja
gradio-app.py c9fffe8b0b Gradio-App 7 hónapja

README.md

Ideas: NotebookLLama

Steps:
Path:

  1. Decide the Topic
  2. Upload a PDF OR - Put in a topic -> Scraped

  3. Report written

  4. 2 Agents debate/interact? Podcast style -> Write a Transcript

  5. TTS Engine (E25 or ) Make the podcast

Instructions:

Running 1B-Model: python 1B-chat-start.py --temperature 0.7 --top_p 0.9 --system_message "you are acting as an old angry uncle and will debate why LLMs are bad" --user_message "I love LLMs"

Running Debator: python 1B-debating-script.py --initial_topic "The future of space exploration" --system_prompt1 "You are an enthusiastic advocate for space exploration" --system_prompt2 "You are a skeptic who believes we should focus on Earth's problems first" --n_turns 4 --temperature 0.8 --top_p 0.9 --model_name "meta-llama/Llama-3.2-1B-Instruct"

Scratch-pad/Running Notes:

Actually this IS THE MOST CONSISTENT PROMPT: Small:

description = """
Laura's voice is expressive and dramatic in delivery, speaking at a fast pace with a very close recording that almost has no background noise.
"""

Large:

description = """
Alisa's voice is consistent, quite expressive and dramatic in delivery, with a very close recording that almost has no background noise.
"""

Small:

description = """
Jenna's voice is consistent, quite expressive and dramatic in delivery, with a very close recording that almost has no background noise.
"""

Bark is cool but just v6 works great, I tried v9 but its quite robotic and that is sad.

So Parler is next-its quite cool for prompting

xTTS-v2 by coquai is cool, however-need to check the license-I think an example is allowed

Torotoise is blocking because it needs HF version that doesnt work with llama-3.2 models so I will probably need to make a seperate env-need to eval if its worth it

Side note: The TTS library is a really cool effort!

Bark-Tests: Best results for speaker/v6 are at speech_output = model.generate(**inputs, temperature = 0.9, semantic_temperature = 0.8) Audio(speech_output[0].cpu().numpy(), rate=sampling_rate)

Tested sound effects:

  • Laugh is probably most effective
  • Sigh is hit or miss
  • Gasps doesn't work
  • A singly hypen is effective
  • Captilisation makes it louder

Ignore/Delete this in final stages, right now this is a "vibe-check" for TTS model(s):

Starting with: Bark but if it falls apart, here is the order

Vibe check:

Higher Barrier to testing (In other words-I was too lazy to test):

Try later:

Resources I used to learn about Suno: