Alternatives
Products that do what 80% faster, 50% less memory, 0% loss of accuracy Llama finetuning does
Hi HN! I'm just sharing a project I've been working on during the LLM Efficiency Challenge - you can now finetune Llama with QLoRA 5x faster than Huggingface's original implementation on your own local GPU. Some highlights: 1. Manual autograd engine - hand derived backprop steps. 2. QLoRA / LoRA 80% faster, 50% less memory. 3. All kernels written in OpenAI's Triton language. 4. 0% loss in accuracy - no approximation methods - all exact. 5. No change of hardware necessary. Supports NVIDIA GPUs since 2018+. CUDA 7.5+. 6. Flash Attention support via Xformers. 7. Supports 4bit and 16bit…
- 1FL
I've been playing around with https://github.com/zphang/minimal-llama/ and https://github.com/tloen/alpaca-lora/blob/main/finetune.py, and wanted to create a simple UI where you can just paste text, tweak the parameters, and finetune the model quickly using a modern GPU. To prepare the data, simply separate your text with two blank lines. There's an inference tab, so you can test how the tuned model behaves. This is my first foray into the world of LLM finetuning, Python, Torch, Transformers, LoRA, PEFT, and Gradio. Enjoy!
2023 · github.com
- 2TL
Hey HN, we wanted to share our repo where we fine-tuned Llama 3.1 on Google TPUs. We’re building AI infra to fine-tune and serve LLMs on non-NVIDIA GPUs (TPUs, Trainium, AMD GPUs). The problem: Right now, 90% of LLM workloads run on NVIDIA GPUs, but there are equally powerful and more cost-effective alternatives out there. For example, training and serving Llama 3.1 on Google TPUs is about 30% cheaper than NVIDIA GPUs. But developer tooling for non-NVIDIA chipsets is lacking. We felt this pain ourselves. We initially tried using PyTorch XLA to train Llama 3.1 on TPUs, but it was rough: xla…
2024 · github.com
- 3
- 4SU
Here's a project I've been working on for the last few months. It's a new (I think) algorithm, that allows to adjust smoothly - and in real time - how many calculations you'd like to do during inference of an LLM model. It seems that it's possible to do just 20-25% of weight multiplications instead of all of them, and still get good inference results. I implemented it to run on M1/M2/M3 GPU. The mmul approximation itself can be pushed to run 2x fast before the quality of output collapses. The inference speed is just a bit faster than Llama.cpp's, because the rest of implementation…
2024 · asciinema.org
- 5L3
Hi everyone, I'm kinda involved in some retrogaming and with some experiments I ran into the following question: "It would be possible to run transformer models bypassing the cpu/ram, connecting the gpu to the nvme?" This is the result of that question itself and some weekend vibecoding (it has the linked library repository in the readme as well), it seems to work, even on consumer gpus, it should work better on professional ones tho
Feb 2026 · github.com
- 6IV
The video demo runs a 7b Model on a normal gaming GPU. I think it already works quite well (accounting for the limited hardware power). :)
2024 · github.com
- 7
- 8
- 9WM
Try it out! https://glhf.chat/ Hey HN! We’ve been working for the past few months on a website to let you easily run (almost) any open-source LLM on autoscaling GPU clusters. It’s free for now while we figure out how to price it, but we expect to be cheaper than most GPU offerings since we can run the models multi-tenant. Unlike Together AI, Fireworks, etc, we’ll run any model that the open-source vLLM project supports: we don’t have a hardcoded list. If you want a specific model or finetune, you don’t have to ask us for it: you can just paste the Hugging Face link in and…
2024 · glhf.chat
- 10KR
I discovered that in LLM inference, keys and values in the KV cache have very different quantization sensitivities. Keys need higher precision than values to maintain quality. I patched llama.cpp to enable different bit-widths for keys vs. values on Apple Silicon. The results are surprising: - K8V4 (8-bit keys, 4-bit values): 59% memory reduction with only 0.86% perplexity loss - K4V8 (4-bit keys, 8-bit values): 59% memory reduction but 6.06% perplexity loss - The configurations use the same number of bits, but K8V4 is 7× better for quality This means you can run LLMs with 2-3× longer…
2025 · github.com
- 11
- 12
Fine-tune LLMs from one YAML. Layer streaming trains an 8B model on a 4 GB laptop GPU. - MakazhanAlpamys/Soup
Aug 2026 · github.com
- 13RQ
Sep 2025 · github.com
- 14FL
2024 · colab.research.google.com
- 15
- 16TV
May 2026 · github.com
- 17LD
2023 · github.com
- 18AT
I recently built a small open-source tool to benchmark different LLM API endpoints — including OpenAI, Claude, and self-hosted models (like llama.cpp). It runs a configurable number of test requests and reports two key metrics: • First-token latency (ms): How long it takes for the first token to appear • Output speed (tokens/sec): Overall output fluency Demo: https://llmapitest.com/ Code: https://github.com/qjr87/llm-api-test The goal is to provide a simple, visual, and reproducible way to evaluate performance across different LLM providers, including…
2025 · llmapitest.com
- 19
- 20
I built slotstream, a way to run Qwen3.8-Flash-Next 4-bit on a low-memory mac starting from 16GB, a 125B parameter model that would need 100GB+ memory/RAM, thanks to expert-offloading/ssd-streaming. Easy to install/update, and mac-native using MLX and Swift. It ships with auto-mode, which makes a good tradeoff between memory usage and speed. I'll be implementing and porting the MTP module for speculative decoding next Local models really are the future of computing!
5d ago · github.com
- 21
- 22L2
Hi all, today we're excited to launch LoraLand: 25 fine-tuned mistral-7b models that outperform #gpt4 on task-specific applications ranging from sentiment detection to question answering. All 25 fine-tuned models… - Outperform GPT-4, GPT-3.5-turbo, and mistral-7b-instruct for specific tasks - Are cost-effectively served from a single GPU through LoRAX - Were trained for less than $8 each on average You can prompt all of the fine-tuned models today and compare their results to mistral-7b-instruct in real time! We'd love to hear comments and feedback from the community
2024 · predibase.com
- 23LF
100% bootstrapped new startup. It lets you fine tune Mistral-7B and SDXL. In particular, for the LLM fine tuning we implemented a dataprep pipeline that turns websites/pdfs/doc files into question-answer pairs for training the small LLM using an big LLM. It includes a GPU scheduler that can do finegrained GPU memory scheduling (Kubernetes can only do whole-GPU, we do it per-GB of GPU memory to pack both inference and fine tuning jobs into the same fleet) to fit model instances into GPU memory to optimally trade off user facing latency with GPU memory utilization It's a pretty…
2023 · docs.helix.ml
- 24RA
Hi there, looking for feedback on my new project "Featherless.AI" The idea is to allow users to run all the models on hugging face instantly. Via the OpenAI API compatible endpoint. Why? Because its a real chore to download models and spin up GPUs, especially if you want to test multiple models. Not to mention GPUs cost multiple dollars an hour to rent. And if we want more people to use open source AI, we got to make it easier for them to try and play with all of them. So what if instead of spinning up dedicated GPUs per model (which is what every provider is doing) We can startup a LLM…
2024 · featherless.ai
Ranked by how close each launch is in meaning, then by votes. Refine with a description →