Search found 20 matches

Author Message

leaf

Post 27-Sep-2024 00:23

Cloned from https://huggingface.co/meta-llama/Llama-3.1-8B-Instruct.

keepthepace

Post 25-Jul-2024 09:17

Looks like mine is getting speed and seeders. Yes, probably better to keep just one.

NotFirst

Post 24-Jul-2024 22:50

I cannot remove mine. I'm seeding both right now.

keepthepace

Post 24-Jul-2024 09:40

Yes, I had stupidly switched off my PC after putting the torrent online :-(
Not sure how to deal with that actually. If your torrent reaches 4 seeders before mine I'll just remove my post.

NotFirst

Post 23-Jul-2024 19:33

Actually I cannot download yours, the torrent is stalled. I will leave mine seeding for now.

NotFirst

Post 23-Jul-2024 19:30

Nice! You beat me by 15 minutes :) I will delete mine and seed yours.

keepthepace

Post 23-Jul-2024 19:11

Fresh from the Metakery!

NotFirst

Post 23-Jul-2024 19:28

New & noteworthy: 128K context, function calling. I haven't tried those on the quantized version yet.

NotFirst

Post 23-Jul-2024 19:26

Llama 3.1 8B Instruct released today by Meta.
GGUF by https://huggingface.co/MaziyarPanahi/Meta-Llama-3.1-8B-Instruct-GGUF
Model Information
The Meta Llama 3.1 collection of multilingual large language models (LLMs) is a collection of pretrained and instruction tuned generative models in 8B, 70B and 405B sizes (text in/text out). The Llama 3.1 instruction tuned text only models (8B, 70B, 405B) are optimized for multilingual dialogue use cases and outperform many of the available open source and closed chat models on common industry benchmarks.
Model developer: Meta
Model Architecture: Llama 3.1 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety.
Supported languages: English, German, French, Italian, Portuguese, Hindi, Spanish, and Thai.
Llama 3.1 family of models. Token counts refer to pretraining data only. All model versions use Grouped-Query Attention (GQA) for improved inference scalability.
Model Release Date: July 23, 2024.
Model card: https://huggingface.co/meta-llama/Meta-Llama-3.1-8B-Instruct

keepthepace

Post 21-Jul-2024 09:39

Q8 of OpenHermes-Llama3 model. Excellent coding capabilities for a 8B model.

NotFirst

Post 22-Jun-2024 02:44

8B q6
https://huggingface.co/SicariusSicariiStuff/LLAMA-3...igned_Alpha_GGUF
!1st

Dreamertist

Post 10-Jun-2024 19:11

I don't really expect L1, L2 or any of the depreciated model torrents to get much attention. They're there for archival reasons if anyone wants to put a bunch of old models up

BoloniniD

Post 10-Jun-2024 12:51

Instruct version of Llama3 8B converted to EXL2 at 8.0bpw precision

Arti

Post 10-Jun-2024 12:41

then seem to be a issue.
this isnt my first time using torrent but first time seeding my file. i put file in my phone and opened torrent file, then choiced destination where my llm file was. FLUD app start checking to 100% and start seeding.
when i opened torrent file in my pc it just didnt download anything, eventhough i was seeding with my phone.

Arti

Post 10-Jun-2024 12:05

this is uncensored version of llama 3 8B 4 bit
keep it safe.
due to internet cost, i can only upload in 2:00 AM to 6:00 AM UTC+03:30 .
i'm uploading with mobile network but try to keep seeding for at least 5-10 person.
if you want to download from hugging face you can download or seed faster you can get this file from here too:
https://huggingface.co/Orenguteng/Llama-3-8B-Lexi-U...ored_Q4_K_M.gguf
i just have this version but you can get download and seed other quantizations too.
thanks :aa:
 

Current time is: 11-Nov 17:43

All times are UTC + 1

Cover our bill…

Donate at ko-fi.com

Or use crypto:

We accept BTC, XMR, LTC, BCH & DOGE