diff options
| author | Shijie <821898965@qq.com> | 2024-01-19 19:53:13 +0800 |
|---|---|---|
| committer | GitHub <noreply@github.com> | 2024-01-19 13:53:13 +0200 |
| commit | 9b75cb2b3ccbed3df2e14c1202168db3e5145095 (patch) | |
| tree | 1ce9e2b388805fefac29ff6feebb326b78d708c8 /gguf-py | |
| parent | de9a147df14e62f54f879d2d15e6c4793107f4fc (diff) | |
llama : support upcoming Qwen2 (#5037)
Diffstat (limited to 'gguf-py')
| -rw-r--r-- | gguf-py/gguf/constants.py | 16 |
1 files changed, 16 insertions, 0 deletions
diff --git a/gguf-py/gguf/constants.py b/gguf-py/gguf/constants.py index 95c58b41..2d9c33c7 100644 --- a/gguf-py/gguf/constants.py +++ b/gguf-py/gguf/constants.py @@ -97,6 +97,7 @@ class MODEL_ARCH(IntEnum): BLOOM = auto() STABLELM = auto() QWEN = auto() + QWEN2 = auto() PHI2 = auto() PLAMO = auto() CODESHELL = auto() @@ -146,6 +147,7 @@ MODEL_ARCH_NAMES: dict[MODEL_ARCH, str] = { MODEL_ARCH.BLOOM: "bloom", MODEL_ARCH.STABLELM: "stablelm", MODEL_ARCH.QWEN: "qwen", + MODEL_ARCH.QWEN2: "qwen2", MODEL_ARCH.PHI2: "phi2", MODEL_ARCH.PLAMO: "plamo", MODEL_ARCH.CODESHELL: "codeshell", @@ -358,6 +360,20 @@ MODEL_TENSORS: dict[MODEL_ARCH, list[MODEL_TENSOR]] = { MODEL_TENSOR.FFN_DOWN, MODEL_TENSOR.FFN_UP, ], + MODEL_ARCH.QWEN2: [ + MODEL_TENSOR.TOKEN_EMBD, + MODEL_TENSOR.OUTPUT_NORM, + MODEL_TENSOR.OUTPUT, + MODEL_TENSOR.ATTN_NORM, + MODEL_TENSOR.ATTN_Q, + MODEL_TENSOR.ATTN_K, + MODEL_TENSOR.ATTN_V, + MODEL_TENSOR.ATTN_OUT, + MODEL_TENSOR.FFN_NORM, + MODEL_TENSOR.FFN_GATE, + MODEL_TENSOR.FFN_DOWN, + MODEL_TENSOR.FFN_UP, + ], MODEL_ARCH.PLAMO: [ MODEL_TENSOR.TOKEN_EMBD, MODEL_TENSOR.OUTPUT_NORM, |