google/gemma-3-4b-it · Discussions

#51 opened 9 days ago by

YovelRVeritex

Update config.json

#50 opened 13 days ago by

olegshulyakov

Update README.md

#49 opened about 1 month ago by

VANNVISAL

Tokenizer's `model_max_length` unusual value

#48 opened about 1 month ago by

awni

Error in loading Model

#46 opened about 1 month ago by

sans07

gemma3-1b-it

#45 opened about 2 months ago by

TAENNY

Explicitly include num_attention_heads and num_key_value_heads in text_config for Gemma3 4B

#44 opened about 2 months ago by

ebsmothers

Deployment issues on sagemaker

#43 opened about 2 months ago by

tychen6677

Please open-source Gemini 1.5 Flash

👍 1

#42 opened about 2 months ago by

drguolai

CUDA error: device-side assert triggered

#41 opened about 2 months ago by

arshiasoori

Low GPU Utilization during inference?

#39 opened about 2 months ago by

BagelBig

Object detection capabilities ?

#38 opened about 2 months ago by

syrineM

SigLIP or SigLIP2 encoder?

6

#37 opened 2 months ago by

orrzohar

text-generation

#36 opened 2 months ago by

rakmik

I'm releasing the speech version of Gemma-3!

👍 ❤️ 5

8

#35 opened 2 months ago by

junnei

Update README.md

#34 opened 2 months ago by

sachin7yadava

Issue with vLLM Deployment of gemma-3-4b-it on Tesla T4 - No Output

#33 opened 2 months ago by

twodaix

Batch processing on a GPU?

3

#32 opened 2 months ago by

buckeye17-bah

Please release also Gemini Flash 1.5 weights.

#31 opened 2 months ago by

ZeroWw

Finetuning script using HuggingFace

5

#30 opened 2 months ago by

2U1

VRAM not freed during long generations (Gemma, max_new_tokens=3000)

#29 opened 2 months ago by

Nessit

The checkpoint you are trying to load has model type `gemma3` but Transformers does not recognize this architecture

#28 opened 2 months ago by

bajajrahul03

Does not works with bitsandbytes 4bit and 8bit

#27 opened 2 months ago by

zokica

Pan&Scan is desable by default ?

3

#26 opened 2 months ago by

Jack-HR-Red

Upload tokenizer.model

#23 opened 3 months ago by

Nexesenex

Having a hard time fine tuning with LoRA

#21 opened 3 months ago by

abcampbell

Adding "vocab_size" to config.json

#20 opened 3 months ago by

nmysore

The checkpoint you are trying to load has model type `gemma3` but Transformers does not recognize this architecture.

#19 opened 3 months ago by

bedoonraj

getting key error while running inference for pipeline api given in "Usage section"

#18 opened 3 months ago by

himanshu-bluebrain

How can I run the pipeline without the Image input?

👍 5

8

#17 opened 3 months ago by

prakhaaaas

AttributeError: 'HybridCache' object has no attribute 'float'

7

#15 opened 3 months ago by

naruto-soop

Not enough parameters in 4b config.json

#14 opened 3 months ago by

krammnic

AttributeError: 'Gemma3Config' object has no attribute 'vocab_size'

➕ 7

8

#13 opened 3 months ago by

MinecAnton209

Add video inference

#12 opened 3 months ago by

merve

Local Installation Video and Testing - Step by Step

#10 opened 3 months ago by

fahdmirzac

Question About Benchmark Version in README

#9 opened 3 months ago by

lianghsun

Feedback: mildly disappointed there isn't 9B version