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

llama : add jina v2 base code (#7596)

Browse source 
* feat: add changes to handle jina v2 base code

* fix: do not complicate things

* fix: fix the usage of the code model

* fix: fix comments

* fix: fix linting issues

* fix: remove ollama patches

* style : minor

---------

Co-authored-by: Georgi Gerganov <ggerganov@gmail.com>
This commit is contained in:
Joan Fontanals 2024-06-06 09:22:41 +02:00 • committed by GitHub
parent 2d08b7fbb4
commit f5d7b268ec
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
5 changed files with 24 additions and 5 deletions
Download patch file Download diff file Expand all files Collapse all files

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

@ -415,6 +415,7 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = {
MODEL_TENSOR.TOKEN_EMBD,
MODEL_TENSOR.TOKEN_EMBD_NORM,
MODEL_TENSOR.TOKEN_TYPES,
MODEL_TENSOR.ATTN_NORM_2,
MODEL_TENSOR.ATTN_OUT_NORM,
MODEL_TENSOR.ATTN_Q,
MODEL_TENSOR.ATTN_Q_NORM,

3
gguf-py/gguf/tensor_mapping.py
View file

@ -102,6 +102,7 @@ class TensorNameMap:
# Attention norm 2
MODEL_TENSOR.ATTN_NORM_2: (
"transformer.h.{bid}.ln_attn", # falcon40b
"encoder.layer.{bid}.layer_norm_1", # jina-v2-code
),
# Attention query-key-value
@ -311,6 +312,7 @@ class TensorNameMap:
"model.layers.{bid}.mlp.c_proj", # starcoder2
"encoder.layer.{bid}.mlp.wo", # jina-bert-v2
"model.layers.{bid}.residual_mlp.w2", # arctic
"encoder.layer.{bid}.mlp.down_layer", # jina-bert-v2
),
MODEL_TENSOR.FFN_DOWN_EXP: (
@ -350,6 +352,7 @@ class TensorNameMap:
"encoder.layers.{bid}.norm2", # nomic-bert
"transformer.decoder_layer.{bid}.rms_norm_3", # Grok
"encoder.layer.{bid}.mlp.layernorm", # jina-bert-v2
"encoder.layer.{bid}.layer_norm_2" # jina-v2-code
),
MODEL_TENSOR.SSM_IN: (