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

llama : add phi-2 + fix NeoX rope + ggml_mul_mat_set_prec (#4490)

Browse source 
* phi2 implementation

* fix breaking change

* phi-2 : various fixes

* phi-2 : use layer norm eps

* py : whitespaces

* llama : fix meta KV override bug

* convert : phi don't add BOS token

* convert : revert "added_tokens_decoder" change

* phi-2 : scale Q instead of KQ for better precision

* ggml : fix NeoX rope to rotate just first n_dims

* cuda : less diff in the rope_neox kernel

* ggml : add ggml_mul_mat_set_prec

ggml-ci

* Update ggml-cuda.cu

Co-authored-by: slaren <slarengh@gmail.com>

* Update ggml-cuda.cu

Co-authored-by: slaren <slarengh@gmail.com>

* cuda : ggml_cuda_op_mul_mat_cublas support F32 precision

* cuda : remove oboslete comment

---------

Co-authored-by: Ebey Abraham <ebeyabraham@microsoft.com>
Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
Co-authored-by: slaren <slarengh@gmail.com>
This commit is contained in:
Ebey Abraham 2023-12-18 17:27:47 +00:00 • committed by GitHub
parent 3c04bf6da8
commit b9e74f9bca
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
9 changed files with 463 additions and 76 deletions
Download patch file Download diff file Expand all files Collapse all files

22
convert-hf-to-gguf.py
View file

@ -182,6 +182,8 @@ class Model:
return QwenModel
if model_architecture == "MixtralForCausalLM":
return MixtralModel
if model_architecture == "PhiForCausalLM":
return Phi2Model
return Model
def _is_model_safetensors(self) -> bool:
@ -221,6 +223,8 @@ class Model:
return gguf.MODEL_ARCH.QWEN
if arch == "MixtralForCausalLM":
return gguf.MODEL_ARCH.LLAMA
if arch == "PhiForCausalLM":
return gguf.MODEL_ARCH.PHI2
raise NotImplementedError(f'Architecture "{arch}" not supported!')
@ -980,6 +984,24 @@ class QwenModel(Model):
print(f"{new_name}, n_dims = {n_dims}, {old_dtype} --> {data.dtype}")
self.gguf_writer.add_tensor(new_name, data)
class Phi2Model(Model):
def set_gguf_parameters(self):
block_count = self.hparams["n_layer"]
self.gguf_writer.add_name("Phi2")
self.gguf_writer.add_context_length(self.hparams["n_positions"])
self.gguf_writer.add_embedding_length(self.hparams["n_embd"])
self.gguf_writer.add_feed_forward_length(4 * self.hparams["n_embd"])
self.gguf_writer.add_block_count(block_count)
self.gguf_writer.add_head_count(self.hparams["n_head"])
self.gguf_writer.add_head_count_kv(self.hparams["n_head"])
self.gguf_writer.add_layer_norm_eps(self.hparams["layer_norm_epsilon"])
self.gguf_writer.add_rope_dimension_count(self.hparams["rotary_dim"])
self.gguf_writer.add_file_type(self.ftype)
self.gguf_writer.add_add_bos_token(False)
###### CONVERSION LOGIC ######