ver4a/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity

model : Add support for Arcee AI's upcoming AFM model (#14185)

Browse source 
* Add Arcee AFM support

* Add draft update code

* Fix linter and update URL, may still not be final

* Update src/llama-model.cpp

Co-authored-by: Xuan-Son Nguyen <thichthat@gmail.com>

* Remote accidental blank line

---------

Co-authored-by: Xuan-Son Nguyen <thichthat@gmail.com>
This commit is contained in:
Bartowski 2025-06-16 00:04:06 +01:00 • committed by GitHub
parent cd355eda7d
commit d7da8dc83a
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
7 changed files with 235 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

14
convert_hf_to_gguf.py
View file

@ -2020,6 +2020,20 @@ class LlamaModel(TextModel):
raise ValueError(f"Unprocessed experts: {experts}")
@ModelBase.register("ArceeForCausalLM")
class ArceeModel(LlamaModel):
model_arch = gguf.MODEL_ARCH.ARCEE
def set_gguf_parameters(self):
super().set_gguf_parameters()
self._try_set_pooling_type()
rope_scaling = self.hparams.get("rope_scaling") or {}
if rope_scaling.get("rope_type", rope_scaling.get("type")) == "yarn" and "factor" in rope_scaling:
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.YARN)
self.gguf_writer.add_rope_scaling_factor(rope_scaling["factor"])
self.gguf_writer.add_rope_scaling_orig_ctx_len(rope_scaling["original_max_position_embeddings"])
@ModelBase.register(
"LlavaForConditionalGeneration", # pixtral
"Mistral3ForConditionalGeneration", # mistral small 3.1