Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-23 21:17:54 +01:00
Code Issues Packages Projects Releases Wiki Activity

Fix ggml to gguf conversion on Windows (#2733)

Browse Source 
This fixes `RuntimeWarning: overflow encountered in long_scalars`

Credit: anon (not mine)
This commit is contained in:
IgnacioFDM 2023-08-23 06:31:09 -03:00 committed by GitHub
parent b8ad1b66b2
commit 7f7ddd5002
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 3 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
convert-llama-ggmlv3-to-gguf.py
View File

@ -1,12 +1,10 @@
import sys, struct, math, argparse, warnings
import sys, struct, math, argparse
from pathlib import Path
import numpy as np
import gguf
warnings.filterwarnings('error')
# Note: Does not support GGML_QKK_64
QK_K = 256
# Items here are (block size, type size)
@ -95,7 +93,7 @@ class Tensor:
pad = ((offset + 31) & ~31) - offset
offset += pad
n_elems = np.prod(self.dims)
n_bytes = (n_elems * tysize) // blksize
n_bytes = np.int64(np.int64(n_elems) * np.int64(tysize)) // np.int64(blksize)
self.start_offset = offset
self.len_bytes = n_bytes
offset += n_bytes
@ -327,11 +325,7 @@ def main():
data = np.memmap(cfg.input, mode = 'r')
model = GGMLV3Model()
print('* Scanning GGML input file')
try:
offset = model.load(data, 0)
except OverflowError:
print(f'!!! Caught overflow loading tensors. The most likely issue is running on Windows but not in WSL. Try running in WSL if possible.', file = sys.stderr)
raise
offset = model.load(data, 0)
print(f'* GGML model hyperparameters: {model.hyperparameters}')
vocab_override = None
params_override = None