Ah sorry, I meant M1/M2 chips (not specifically M1/2 CPUs). As far as I know, the 4-bit NormalFloat format that is used in QLoRA is currently only supported on Nvidia GPUs (
https://
github.com/TimDettmers/bi
tsandbytes/issues/485
…). Maybe the repo you mentioned uses a different type of quantized training.
OPEN SOURCE
-
QLoRA 4-bit NormalFloat format supported only on Nvidia GPUs
By
–
-
LoRA Compared to Llama-Adapter and Llama-Adapter v2
By
–
LoRA is a parameter-efficient finetuning technique, yes. I recently compared to Llama-Adapter and Llama-Adapter v2:
-
SeamlessM4T Breakthrough in Multilingual Speech Translation
By
–
SeamlessM4T represents a significant breakthrough in the field of speech-to-speech & speech-to-text by addressing the challenges of limited language coverage & a reliance on separate systems. More details https://
bit.ly/45g2pMq -
Fine-tuning LLaMA2 Workshop: One-Day In-Person Event
By
–
If you want to explore finetuning LLaMA2, we'll be talking at and helping out with a 1 day in-person event focused explicitly on finetuning OSS models Hopefully this guide will come in handy! RSVP here (s/o @swyx and @NaderLikeLadder for organizing): https://
partiful.com/e/T4ngRPaU2uUT
XM8pN17d
… -
Flowise v1.3.4 Release: VectorDB and SQLChain Enhancements
By
–
It's release time again Flowise v1.3.4 is now freshly baked with: VectorDB Document Loader Chat history variable 3 new vector stores: Milvus, PgVector, Zep SQLChain now supports MySQL, MsSQL, PostgreSQL Serp API Loader
-
Lit-GPT: Lightning AI’s New GitHub Repository Launch
By
–
Lit-GPT 🙂 https://
github.com/Lightning-AI/l
it-gpt
… -
SAS Developer Portal Launches Open-Source Tools for Developers
By
–
Developers, we've got you . The SAS Developer Portal was designed with you in mind. See it in action with Joe Furbee, SAS Developer Advocate at #ExploreSAS. We have the open-source tools to make you even more successful in your field http://
2.sas.com/6016Poyhq #Analytics -
LoRA Services vs Open-Source: Llama 2 Competitive Comparison
By
–
Will be interesting how this LoRA-on-demand service will compare to open-source LoRA on prem. Here's a little reminder that open-source Llama 2 compares very favorably to ChatGPT / GPT 3.5
-
ML Evaluation Metrics and Lit-GPT Evaluation Framework
By
–
I have a write-up explaining BLEU, ROUGE, and BERTScore in Chapter 19 of my ML Q and AI book: https://
leanpub.com/machine-learni
ng-q-and-ai/
… Other than that, Lit-GPT has currently the Eleuther AI Evaluation Harness implemented: https://
github.com/Lightning-AI/l
it-gpt/blob/main/tutorials/evaluation.md
… -
Supported Models Toggle via Checkpoint Directory Flag
By
–
Sure, the same code works for all supported models. You can toggle these via the –checkpoint_dir flag pointing to the directory where you downloaded the model. Here's a list of the currently supported models: