Skip to content

Commit 01f2224

Browse files
committed
falcon : write file type
1 parent 38b16df commit 01f2224

File tree

2 files changed

+2
-1
lines changed

2 files changed

+2
-1
lines changed

convert-falcon-hf-to-gguf.py

Lines changed: 1 addition & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -107,6 +107,7 @@ def count_model_parts(dir_model: str) -> int:
107107
else:
108108
gguf_writer.add_head_count_kv(1)
109109
gguf_writer.add_layer_norm_eps(hparams["layer_norm_epsilon"])
110+
gguf_writer.add_file_type(ftype)
110111

111112
# TOKENIZATION
112113

scripts/run-all-ppl.sh

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
#!/bin/bash
22

33
qnt=(f16 q8_0 q6_k q5_k q5_1 q5_0 q4_k q4_1 q4_0 q3_k q2_k)
4-
args="--no-mmap -ngl 999 -t 8"
4+
args="-ngl 999 -t 8"
55

66
if [ -z "$1" ]; then
77
echo "usage: $0 <model> [qnt] [args]"

0 commit comments

Comments
 (0)