Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-26 14:20:31 +01:00
Code Issues Packages Projects Releases Wiki Activity

llama : add 128k yarn context for Qwen (#10698)

Browse Source 
* add 128k yarn context for Qwen

* added property for model tensors

* removing useless line
This commit is contained in:
Robert Collins 2024-12-07 16:12:27 -05:00 committed by GitHub
parent 3573fa8e7b
commit 62e84d9848
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
2 changed files with 9 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
convert_hf_to_gguf.py
View File

@ -1992,6 +1992,14 @@ class Qwen2Model(Model):
except FileNotFoundError: except FileNotFoundError:
self._set_vocab_gpt2() self._set_vocab_gpt2()
def set_gguf_parameters(self):
super().set_gguf_parameters()
if self.hparams.get("rope_scaling") is not None and "factor" in self.hparams["rope_scaling"]:
if self.hparams["rope_scaling"].get("type") == "yarn":
self.gguf_writer.add_rope_scaling_type(gguf.RopeScalingType.YARN)
self.gguf_writer.add_rope_scaling_factor(self.hparams["rope_scaling"]["factor"])
self.gguf_writer.add_rope_scaling_orig_ctx_len(self.hparams["rope_scaling"]["original_max_position_embeddings"])
@Model.register("Qwen2MoeForCausalLM") @Model.register("Qwen2MoeForCausalLM")
class Qwen2MoeModel(Model): class Qwen2MoeModel(Model):

1
gguf-py/gguf/constants.py
View File

@ -761,6 +761,7 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.OUTPUT_NORM, MODEL_TENSOR.OUTPUT_NORM,
MODEL_TENSOR.OUTPUT, MODEL_TENSOR.OUTPUT,
MODEL_TENSOR.ROPE_FREQS,
MODEL_TENSOR.ATTN_NORM, MODEL_TENSOR.ATTN_NORM,
MODEL_TENSOR.ATTN_Q, MODEL_TENSOR.ATTN_Q,
MODEL_TENSOR.ATTN_K, MODEL_TENSOR.ATTN_K,