潘其威(William)
15db2cdc44
Update _base.py
...
fix problem that recursively adding file extension to model_base_name
2023-05-30 07:26:42 +08:00
潘其威(William)
cfa7271617
Update _base.py
...
fix variable not found error
2023-05-30 07:22:10 +08:00
潘其威(William)
e5771fb206
Update _base.py
...
fix key mismatch
2023-05-30 06:44:45 +08:00
潘其威(William)
61a4ea035f
Update auto.py
...
add back save_dir for backward compatible
2023-05-30 06:43:00 +08:00
潘其威(William)
ea74e15199
Update _base.py
...
add model_name_or_path and model_file_base_name to BaseQuantizeConfig for better model file management; add back save_dir to .from_quantized() for backward compatible
2023-05-30 06:40:31 +08:00
TheBloke
b7bb50b4d5
Fix bug added after merge
2023-05-25 07:05:51 +01:00
Tom Jobbins
492255b400
Merge branch 'main' into TheBloke_support-HF-download
2023-05-25 07:02:13 +01:00
PanQiWei
94ef4d5ada
update basic usage example code
2023-05-24 17:56:46 +08:00
PanQiWei
c89bb6450c
correct typo of function name
2023-05-24 17:43:38 +08:00
PanQiWei
10347fdd7b
remove full_cpu_offload argument and unify model dispatch strategy
2023-05-24 17:41:04 +08:00
PanQiWei
379f24c2a5
remove add_align_logits_hook_to_model
2023-05-24 17:01:57 +08:00
PanQiWei
749dba1a7e
disable add_align_logits_hook_to_model for now
2023-05-24 13:42:06 +08:00
PanQiWei
58c1b509f0
support add_align_logits_hook_to_model
2023-05-24 12:50:30 +08:00
PanQiWei
21ab7c435a
make comments more readable
2023-05-24 11:38:29 +08:00
PanQiWei
c31b370228
make_sure_not_tensor_in_meta_device before load checkpoint
2023-05-24 11:32:45 +08:00
PanQiWei
63f1b4e073
remove comment
2023-05-24 11:23:07 +08:00
PanQiWei
057c39e3f2
fix meta device bug when use low_cpu_mem_usage
2023-05-24 11:19:59 +08:00
PanQiWei
e2e7809a1f
always to enable QuantLinear bias to make compatible with model quantized from other frameworks
2023-05-24 10:56:31 +08:00
PanQiWei
191da8141e
fix device mismatch
2023-05-23 23:22:52 +08:00
PanQiWei
e4e90e8b0a
add warmup_triton method
2023-05-23 23:18:46 +08:00
PanQiWei
ed14d3a786
fix save quantized model failed when load pretrained model using CPU offload
2023-05-23 23:17:11 +08:00
PanQiWei
6476ee4235
add options: 'low_cpu_mem_usage' and 'full_cpu_offload'
2023-05-23 22:51:00 +08:00
PanQiWei
1b2159bd4c
add more help functions
2023-05-23 19:30:28 +08:00
TheBloke
bf633c298e
Clean up some unused params
2023-05-20 10:32:27 +01:00
PanQiWei
86b3b52c63
fix ImportError when triton is not installed
2023-05-20 16:15:20 +08:00
潘其威(William)
13defe253a
Merge pull request #84 from TheBloke/TheBloke_forward-positional-args
...
Forward position args to allow `model(tokens)` syntax
2023-05-20 15:04:27 +08:00
潘其威(William)
1ef0af824a
Merge pull request #80 from PanQiWei/user_customized_device_map
...
Support users customize `device_map`
2023-05-20 15:00:05 +08:00
TheBloke
e5c8479100
Remove debugging print line
2023-05-19 17:50:48 +01:00
TheBloke
735f7df4cc
Add push_to_hub for HF hub uploading
2023-05-19 17:10:57 +01:00
TheBloke
908b338436
Initial support for model loading from HF hub
2023-05-19 15:57:05 +01:00
TheBloke
a397f00cc3
Implement HF cached download for quantize_config
2023-05-19 15:15:43 +01:00
TheBloke
7f165337ed
Forward position args to allow syntax
2023-05-16 12:19:52 +01:00
PanQiWei
759d6953d4
support user customize device_map
2023-05-15 13:26:38 +08:00
PanQiWei
07e06fa08c
make compatible with older transformers version
2023-05-15 13:26:18 +08:00
oobabooga
86c7021285
Look for .pt files
2023-05-15 00:00:05 -03:00
PanQiWei
d5429441ef
add GPTJ fused attention module
2023-05-14 16:17:21 +08:00
PanQiWei
5445c67190
add library version comparison help functions
2023-05-14 16:16:06 +08:00
PanQiWei
de33d26d67
fix bugs
2023-05-14 13:07:18 +08:00
PanQiWei
2273f9ef39
refactor file structure for triton kernels
2023-05-14 11:49:10 +08:00
PanQiWei
fef1a4fe4b
make code clean and extendable
2023-05-12 20:11:55 +08:00
PanQiWei
d718d63e9c
add import_utils.py for commonly used module importation
2023-05-12 19:58:48 +08:00
PanQiWei
c5ff195764
skip fused module injection instead of raising error if it's not supported yet.
2023-05-12 19:36:00 +08:00
PanQiWei
f159aeabb6
refactor .from_quantized api and improve model loading strategy
2023-05-12 18:09:50 +08:00
PanQiWei
4bb10fda49
groupsize -> group_size
2023-05-12 13:37:52 +08:00
qwopqwop200
3ff6ab18cb
Merge branch 'main' into faster-llama
2023-05-06 00:20:29 +09:00
TheBloke
1b3329b399
Fix 'groupsize' -> 'group_size' in all other .py files. I haven't touched any CUDA kernels in case there's any complexity there I don't understand
2023-05-05 14:44:16 +01:00
TheBloke
f61ce12271
Change 'groupsize' to 'group_size' everywhere. Turns out this is easier than 'groupsize' due to dependencies in other files.
2023-05-05 13:36:00 +01:00
TheBloke
f64c71e779
Change referenes to 'group_size' to 'groupsize' to match rest of this file
2023-05-05 13:21:13 +01:00
PanQiWei
1c6bb69fae
fix attribute name error
2023-05-04 22:10:33 +08:00
潘其威(William)
771b650a7c
Merge pull request #38 from PanQiWei/faster-cuda-no-actorder
...
Faster cuda no actorder
2023-05-04 21:47:19 +08:00
