From: Farbod Bijary <redacted>
Date: Fri, 16 Aug 2024 10:36:30 +0000 (+0330)
Subject: py : fix wrong input type for raw_dtype in ggml to gguf scripts (#8928)
X-Git-Tag: upstream/0.0.4488~890
X-Git-Url: https://git.djapps.eu/?a=commitdiff_plain;h=ee2984bdaf10c14d440ad873a049bcc09b786d9b;p=pkg%2Fggml%2Fsources%2Fllama.cpp

py : fix wrong input type for raw_dtype in ggml to gguf scripts (#8928)

Co-authored-by: farbod <redacted>
---

diff --git a/convert_llama_ggml_to_gguf.py b/convert_llama_ggml_to_gguf.py
index 7b00b439..29b14e98 100755
--- a/convert_llama_ggml_to_gguf.py
+++ b/convert_llama_ggml_to_gguf.py
@@ -116,7 +116,7 @@ class Tensor:
         assert quant is not None, 'Unknown tensor type'
         (blksize, tysize) = quant
         offset += 12
-        self.dtype= dtype
+        self.dtype= gguf.GGMLQuantizationType(dtype)
         self.dims = struct.unpack(f'<{n_dims}I', data[offset:offset + (4 * n_dims)])
         offset += 4 * n_dims
         self.name = bytes(data[offset:offset + name_len])