Woosuk Kwon
|
e9d3f2ff77
|
Add memory analyzer & utomatically configure KV cache size (#6)
|
2023-03-11 23:23:14 -08:00 |
|
Woosuk Kwon
|
1a7eb7da61
|
Support beam search & parallel generation (#7)
|
2023-03-10 09:58:21 -08:00 |
|
Woosuk Kwon
|
04e5acc08e
|
Fix a bug in 1D input shape (#5)
|
2023-03-06 10:05:27 -08:00 |
|
Woosuk Kwon
|
3e9f991d6a
|
Use FlashAttention for multi_query_kv_attention (#4)
|
2023-03-01 21:13:08 -08:00 |
|
Woosuk Kwon
|
0deacbce6e
|
Implement single_query_cached_kv_attention kernel (#3)
|
2023-03-01 15:02:19 -08:00 |
|
Woosuk Kwon
|
cbf8779afa
|
Fix a bug in tying OPT embeddings (#1)
|
2023-02-24 16:29:36 -08:00 |
|
Woosuk Kwon
|
6aef2278f4
|
[Minor] Fix printing format
|
2023-02-24 11:56:06 +00:00 |
|
Woosuk Kwon
|
1132fae0ca
|
Add Frontend
|
2023-02-24 11:46:43 +00:00 |
|
Woosuk Kwon
|
46ce1356f7
|
Add max_num_steps to SamplingParams
|
2023-02-24 11:44:40 +00:00 |
|
Woosuk Kwon
|
b39f149a08
|
Add is_finished
|
2023-02-24 11:44:21 +00:00 |
|
Woosuk Kwon
|
ef6098ec51
|
Merge pre_step and step
|
2023-02-24 10:36:08 +00:00 |
|
Woosuk Kwon
|
53f70e7334
|
Reduce the number of states in scheduler
|
2023-02-24 10:22:39 +00:00 |
|
Woosuk Kwon
|
762fd1c3fa
|
Refactor and annotate types for attention
|
2023-02-24 08:58:46 +00:00 |
|
Woosuk Kwon
|
7f22f90e8c
|
Remove xformers
|
2023-02-24 08:36:16 +00:00 |
|
Woosuk Kwon
|
afdbe5d373
|
[WIP] Add server script
|
2023-02-24 01:33:37 +00:00 |
|
Woosuk Kwon
|
932844f1cd
|
Fix attention
|
2023-02-23 23:02:25 +00:00 |
|
Woosuk Kwon
|
ba84b8728a
|
Fix attention
|
2023-02-23 22:29:46 +00:00 |
|
Woosuk Kwon
|
87e0bcd426
|
Fix attention
|
2023-02-23 21:32:02 +00:00 |
|
Woosuk Kwon
|
1ce1333573
|
Set default dtype to half
|
2023-02-23 21:31:39 +00:00 |
|
Woosuk Kwon
|
de0fabbc5c
|
Fix sampler
|
2023-02-23 20:30:12 +00:00 |
|
Woosuk Kwon
|
fdd0f2f472
|
Minor
|
2023-02-23 20:23:47 +00:00 |
|
Woosuk Kwon
|
7f985166f7
|
Consider pempty tensor
|
2023-02-23 20:20:33 +00:00 |
|
Woosuk Kwon
|
86f9eb6d39
|
Fix typo
|
2023-02-23 20:19:24 +00:00 |
|
Woosuk Kwon
|
1f6c7ef437
|
Add controller
|
2023-02-23 09:32:19 +00:00 |
|
Woosuk Kwon
|
d4bc1a4d24
|
Add unoptimized OPT Attention
|
2023-02-23 09:31:55 +00:00 |
|
Woosuk Kwon
|
b56b6ca0d6
|
Add greedy sampler
|
2023-02-23 09:26:09 +00:00 |
|
Woosuk Kwon
|
343cea3dbc
|
Add seq_ids to input metadata
|
2023-02-23 09:25:01 +00:00 |
|
Woosuk Kwon
|
4f6f4967f6
|
Add get_block_table
|
2023-02-23 07:55:14 +00:00 |
|
Woosuk Kwon
|
331fa0b042
|
Implement scheduler.step & Add a threshold for batch size
|
2023-02-23 07:54:20 +00:00 |
|
Woosuk Kwon
|
501c4bd0cd
|
decoding.py -> sampling_params.py
|
2023-02-23 07:39:20 +00:00 |
|
Woosuk Kwon
|
86c682cd32
|
DecodingParams -> SamplingParams
|
2023-02-23 07:38:43 +00:00 |
|
Woosuk Kwon
|
af16c05074
|
Add get_len
|
2023-02-23 05:58:04 +00:00 |
|
Woosuk Kwon
|
d094512296
|
Move max_context_len
|
2023-02-23 04:57:46 +00:00 |
|
Woosuk Kwon
|
4b1ac23f53
|
Fix slot mapping
|
2023-02-23 00:10:07 +00:00 |
|
Woosuk Kwon
|
8290fce47d
|
Add Worker class
|
2023-02-22 19:01:38 +00:00 |
|
Woosuk Kwon
|
7b6844e590
|
Add input metadata
|
2023-02-22 19:01:20 +00:00 |
|
Woosuk Kwon
|
608f74ffe5
|
Minor
|
2023-02-22 18:08:25 +00:00 |
|
Woosuk Kwon
|
709a69176e
|
Move worker/models -> models
|
2023-02-22 18:03:48 +00:00 |
|
Woosuk Kwon
|
6f058c7ba8
|
Implement cache ops
|
2023-02-16 07:47:03 +00:00 |
|
Woosuk Kwon
|
a1c67e6db8
|
Minor
|
2023-02-16 01:42:53 +00:00 |
|
Woosuk Kwon
|
9e68a6827e
|
Fix return type error
|
2023-02-16 01:33:03 +00:00 |
|
Woosuk Kwon
|
8edcabc737
|
Add warning
|
2023-02-16 01:28:17 +00:00 |
|
Woosuk Kwon
|
2f4887de77
|
Fix KVCache shape
|
2023-02-16 01:24:45 +00:00 |
|
Woosuk Kwon
|
3363c27d19
|
Add __repr__
|
2023-02-14 09:34:07 +00:00 |
|
Woosuk Kwon
|
2729087efe
|
Fix a ref count bug in swapping
|
2023-02-14 09:20:12 +00:00 |
|
Woosuk Kwon
|
c128c2ed03
|
Fix double free
|
2023-02-14 09:14:05 +00:00 |
|
Woosuk Kwon
|
c80ed212a1
|
Minor
|
2023-02-14 09:06:27 +00:00 |
|
Woosuk Kwon
|
263e91b386
|
Minor
|
2023-02-14 09:05:39 +00:00 |
|
Woosuk Kwon
|
be1e2163c9
|
Fix memory leak in swapping
|
2023-02-14 08:46:33 +00:00 |
|
Woosuk Kwon
|
7e5f604e68
|
Fix bugs in scheduler
|
2023-02-14 02:25:32 +00:00 |
|