-
Please add Qwen2 support
```
EETQ_CAUSAL_LM_MODEL_MAP = {
"llama": LlamaEETQForCausalLM,
"baichuan": BaichuanEETQForCausalLM,
"gemma": GemmaEETQForCausalLM
}
```
-
Hello Authors,
Thank you for your incredible work and the comprehensive experiments presented in the paper.
I have a question regarding the implementation of attacks. Specifically, some attacks,…
-
Since the latest models, such as Llama 3 and Gemma, adopt extremely large vocabularies (128-256K), the size of logits can become very large, consuming a large proportion of VRAM. For example, the foll…
-
### Description of the bug:
I ran the Gemema-7B model based on the code in the example, and found that the model's answers were rather poor and didn't seem to understand my question at all. Is this …
-
Edited by PP to add that this issue was noted during splitting an experiment, which turned out to be misannotated so it's not a problem, but in general, when we have FactorValues with multiple stateme…
-
Compile the database of learning materials that contains longer and more reasonable materials.
-
URLs are created relative to settings in `Gemma.properties`, but those are redundant since it is also part of the servlet context configuration. It would also be more robust and work consistently in d…
-
hello. I am getting an error when running the sample below.
The request file does not exist in the original source,
I copied and used the preprocessor_config.json file in the same model family.
…
-
```
[Gemma - http-nio-8181-exec-266 (2022-04-21 11:13:33,376)] ERROR ubic.gemma.core.visualization.ExperimentalDesignVisualizationServiceImpl.sortVectorDataByDesign(121) | Did not find cached layout …
-
**Is your feature request related to a problem? Please describe.**
ShieldGemma was released in Jul 2024 with [v0.14.3](https://github.com/keras-team/keras-hub/releases/tag/v0.14.3), but since then …