Tracked topic
Gemma is a family of open-weight language models released by Google for text generation and related NLP tasks.
… I'm sure the Gemma 4 series will be useful to some people. Not me, though... I just wish larger tech companies would take a page out of OpenAI's book and release actually competitive OSS instead of putting out generic models just to say they support public research. …
… That was a great read, for anyone wanting a deeper understanding of the architecture, read https://huggingface.co/blog/rishiraj/matformer-in-gemma-3n Thanks for day 0 MLX support 🙏 This comment has been hidden Thank you so much for this release! …
… Use LoRA on vision-language projection layers Enable 4-bit quantization to reduce memory usage could you please release the training scripts and deepspeed configuration of 70B model? · Here are all the config used. …
… Indeed we're working close with the AWS SageMaker team on releasing those, we'll ping you once released but should be soon 🤗 EmbeddingGemma looks really promising — especially the focus on efficiency. …
… Replicates on PlanBench Blocksworld and on four open-weights models Qwen3.6-35B-A3B, Gemma-4-26B-A4B-it, gpt-oss-20b, Nemotron-3-Nano-30B-A3B . …