PR #6635 Support StableLM2 12B

StableLM2 12B support for huggingface -> GGUF

ashishdatta committed 2 years ago

StableLM12 tensormapping and constants

ashishdatta committed 2 years ago

StableLM-2-12b model support

ashishdatta committed 2 years ago

fix

ashishdatta committed 2 years ago

Added 12B support

ashishdatta committed 2 years ago

Removed autoformatting; resolved bug where model_arch was not selecting StableLM2

ashishdatta committed 2 years ago

Formatting

ashishdatta committed 2 years ago

Do QK norm stacking in model conversion step

ashishdatta committed 2 years ago

Converge StableLM and StableLM2 code to simplify graph construction

ashishdatta committed 2 years ago

Fix accidental removal

ashishdatta committed 2 years ago

Removed warnings

ashishdatta committed 2 years ago

Revert formatter

ashishdatta committed 2 years ago

Move QK norm stack to private function so it's easier to read

ashishdatta committed 2 years ago

refactor stablelm graph builder to support 1.6, 3b and 12b more efficiently

ashishdatta committed 2 years ago

Proper check for None type for new_name to avoid crash; formatting; revert change to base class `write_tensors()`

ashishdatta committed 2 years ago

Format

ashishdatta committed 2 years ago

Formatting

ashishdatta committed 2 years ago

format

ashishdatta committed 2 years ago

Fix incorrect check for K norm

ashishdatta committed 2 years ago

space after commas; Keep indentation multiple of 4 spaces

ashishdatta committed 2 years ago

Flake8 format

ashishdatta committed 2 years ago

Removed unnecessary conditional branches

ashishdatta committed 2 years ago

Removed unused comment

ashishdatta committed 2 years ago

Fixed incorrect tensor passing

ashishdatta committed 2 years ago

Format

ashishdatta committed 2 years ago

llama.cpp Support StableLM2 12B #6635 Merged

llama.cpp
Support StableLM2 12B
#6635

Merged