mirror of
https://github.com/turboderp-org/exllamav3.git
synced 2026-04-28 18:21:34 +00:00
44 lines
1.3 KiB
JSON
44 lines
1.3 KiB
JSON
[
|
|
{
|
|
"load_fn": "exllamav2",
|
|
"fwd_fn": "exllamav2",
|
|
"label": "EXL2 3.0bpw H6",
|
|
"model_dir": "/mnt/str/eval_models/llama3.1-8b-instruct/exl2/3.0bpw/"
|
|
},
|
|
{
|
|
"load_fn": "exllamav2",
|
|
"fwd_fn": "exllamav2",
|
|
"label": "EXL2 3.5bpw H6",
|
|
"model_dir": "/mnt/str/eval_models/llama3.1-8b-instruct/exl2/3.5bpw/"
|
|
},
|
|
{
|
|
"load_fn": "exllamav2",
|
|
"fwd_fn": "exllamav2",
|
|
"label": "EXL2 4.0bpw H6",
|
|
"model_dir": "/mnt/str/eval_models/llama3.1-8b-instruct/exl2/4.0bpw/"
|
|
},
|
|
{
|
|
"load_fn": "exllamav2",
|
|
"fwd_fn": "exllamav2",
|
|
"label": "EXL2 4.5bpw H6",
|
|
"model_dir": "/mnt/str/eval_models/llama3.1-8b-instruct/exl2/4.5bpw/"
|
|
},
|
|
{
|
|
"load_fn": "exllamav2",
|
|
"fwd_fn": "exllamav2",
|
|
"label": "EXL2 5.0bpw H6",
|
|
"model_dir": "/mnt/str/eval_models/llama3.1-8b-instruct/exl2/5.0bpw/"
|
|
},
|
|
{
|
|
"load_fn": "exllamav2",
|
|
"fwd_fn": "exllamav2",
|
|
"label": "EXL2 6.0bpw H6",
|
|
"model_dir": "/mnt/str/eval_models/llama3.1-8b-instruct/exl2/6.0bpw/"
|
|
},
|
|
{
|
|
"load_fn": "exllamav2",
|
|
"fwd_fn": "exllamav2",
|
|
"label": "EXL2 8.0bpw H6",
|
|
"model_dir": "/mnt/str/eval_models/llama3.1-8b-instruct/exl2/8.0bpw/"
|
|
}
|
|
] |