e74bd41e0f
feat(server): add paged attention to flash models ( #516 )
...
Closes #478
2023-06-30 19:09:59 +02:00
70f485bf9f
feat(router): add header option to disable buffering for the generate_stream response ( #498 )
...
# This PR adds an http header option to disable buffering for the
generate_stream endpoint response stream.
Problem: If a model is run behind a proxy server such as nginx that has
buffering enabled then the response stream from generate_stream gets
aggregated into a single response which basically disables streaming.
Instead of getting a chunked response where each token is presented over
time the response presents everything all at once.
Solution: This change adds the `X-Accel-Buffering` http header which
disables buffering for the generate_stream response, allowing the
response to stream properly.
2023-06-28 11:50:12 +02:00
bd3a9d8e85
fix(router): add timeout on flume sends ( #488 )
2023-06-23 14:58:28 +02:00
f59fb8b630
feat(router): add ngrok integration ( #453 )
2023-06-16 16:25:11 +02:00
19c41824cb
chore: update openapi schema
2023-06-05 18:16:08 +02:00
895c5f1562
feat(server): only compute prefill logprobs when asked ( #406 )
...
Close #288
2023-06-02 17:12:30 +02:00
218c9adaa5
feat: decrease IPC proto size ( #367 )
...
Closes #307 #308
2023-05-24 19:19:57 +02:00
942005386a
feat(router): log input/ouput at debug level ( #364 )
...
@njhill FYI
2023-05-23 20:47:37 +02:00
5a58226130
fix(server): fix decode token ( #334 )
...
Fixes #333
---------
Co-authored-by: Nicolas Patry <patry.nicolas@protonmail.com>
2023-05-16 23:23:27 +02:00
68e9d6ab33
feat(server): shard token decode ( #303 )
2023-05-10 15:48:21 +02:00
e250282213
feat(docker): add benchmarking tool to docker image ( #298 )
2023-05-09 13:19:31 +02:00
926fd9a010
feat(router): Adding response schema for compat_generate ( #292 )
2023-05-09 12:38:09 +02:00
b4fe248b17
fix(launcher): handle hub branches ( #278 )
2023-05-04 15:14:28 +02:00
411b0d4e1f
chore(github): add templates ( #264 )
2023-05-02 15:43:19 +02:00
e86cca9723
Adding docs on how dynamic batching works. ( #258 )
...
This PR starts the minimal possible amount of explanation I could think
of. It tries to explain how dynamic batching occurs, the interactions
with past key values and ignores the padding problem.
Maybe some drawings could help too but I kept it to text for now.
2023-05-01 14:16:50 +02:00
f092ba9b22
feat(server): add watermarking tests ( #248 )
2023-04-27 19:16:35 +02:00
db2b4e0754
feat(router): new healthcheck that skips the queue ( #244 )
...
Co-authored-by: OlivierDehaene <23298448+OlivierDehaene@users.noreply.github.com>
Co-authored-by: OlivierDehaene <olivier@huggingface.co>
2023-04-26 20:23:54 +02:00
c4fb09f2ae
feat(router): add tests to validation ( #237 )
2023-04-26 16:14:40 +02:00
45344244cf
Starting some routing tests. ( #233 )
2023-04-25 14:13:14 +02:00
8b182eb986
feat(router): add endpoint info to /info route ( #228 )
2023-04-25 13:11:18 +02:00
ebc74d5666
feat(router): use number of tokens in batch as input for dynamic batching ( #226 )
...
Co-authored-by: Nick Hill <nickhill@us.ibm.com>
2023-04-24 17:59:00 +02:00
6ded76a4ae
v0.6.0 ( #222 )
2023-04-21 21:00:57 +02:00
343437c7b5
feat(router): add device and dtype info ( #215 )
2023-04-21 15:36:29 +02:00
709d8936f6
feat(router): drop requests when client closes the channel ( #202 )
2023-04-20 11:07:40 +02:00
b6ee0ec7b0
feat(router): add git sha to info route ( #208 )
2023-04-19 21:36:59 +02:00
252f42c1e6
fix(router): add auth token to get model info ( #207 )
2023-04-19 20:06:06 +02:00
2475aede61
feat(router): add info route ( #196 )
...
close #125
2023-04-18 16:16:06 +02:00
c13b9d87c9
fix(router): fix truncation ( #190 )
...
closes #189
2023-04-17 16:51:53 +02:00
64347b05ff
fix(ci): fix CVE in github-slug-action ( #174 )
2023-04-13 12:43:05 +02:00
6f0f1d70f6
v0.5.0 ( #168 )
2023-04-11 20:32:18 +02:00
9987960062
feat(router): make router input validation optional ( #164 )
2023-04-09 20:22:27 +02:00
7dec65a244
fix(router): use buckets for metrics histograms ( #163 )
2023-04-09 20:13:28 +02:00
5cddc055e6
fix(rust-client): use join_all instead of select_all to hopefully fix nccl issues ( #162 )
2023-04-09 20:07:02 +02:00
fef1a1c381
v0.4.3 ( #152 )
2023-03-30 17:28:14 +02:00
84722f3e33
v0.4.2 ( #151 )
2023-03-30 17:10:01 +02:00
610bb1f978
feat(benchmark): tui based benchmarking tool ( #149 )
2023-03-30 15:26:27 +02:00
d503e8f09d
feat: aws sagemaker compatible image ( #147 )
...
The only difference is that now it pushes to
registry.internal.huggingface.tech/api-inference/community/text-generation-inference/sagemaker:...
instead of
registry.internal.huggingface.tech/api-inference/community/text-generation-inference:sagemaker-...
---------
Co-authored-by: Philipp Schmid <32632186+philschmid@users.noreply.github.com>
2023-03-29 21:38:30 +02:00
f000068944
feat(server): clear cache on error ( #143 )
2023-03-28 11:29:35 +02:00
ab5fd8cf93
v0.4.1 ( #140 )
2023-03-26 16:37:51 +02:00
b49dbf2d88
fix(server): use server tokenizer as gt ( #128 )
2023-03-16 12:12:26 +01:00
cbd36aa4d1
fix(server): revert gpt-neox optims ( #123 )
2023-03-13 22:57:08 +01:00
411d6247f4
v0.4.0 ( #119 )
2023-03-09 16:07:01 +01:00
55bd4fed7d
feat(router): add best_of parameter ( #117 )
2023-03-09 15:30:54 +01:00
e8bfe199ba
feat(router): support left truncation ( #115 )
...
closes #111
2023-03-09 13:10:30 +01:00
1a2d68250a
feat: support typical sampling ( #114 )
...
closes #112
2023-03-09 11:33:57 +01:00
3fef90d50f
feat(clients): Python client ( #103 )
2023-03-07 18:52:22 +01:00
cd5961b5da
feat: allow local models ( #101 )
...
closes #99
2023-03-06 14:39:36 +01:00
1c19b0934e
v0.3.2 ( #97 )
2023-03-03 18:42:20 +01:00
9b8ea6a6c7
feat(server): add logits watermark ( #90 )
2023-03-02 12:30:41 +01:00
f874c47831
feat(router): add api-inference headers ( #91 )
2023-03-02 11:41:51 +01:00
4e685d907e
feat(router): ask hf.co for pipelinetag to decide on compat_return_full_text ( #89 )
2023-02-28 10:19:32 +01:00
21340f24ba
feat(router): add legacy route for api-inference support ( #88 )
2023-02-27 14:56:58 +01:00
0ac184ce77
feat(server): add special token bool ( #85 )
2023-02-24 15:55:57 +01:00
4b1c9720c0
v0.3.1 ( #84 )
2023-02-24 13:27:41 +01:00
6796d38c6d
feat(router): add cors allow origin options ( #73 )
2023-02-17 18:22:00 +01:00
c720555adc
v0.3.0 ( #72 )
2023-02-16 17:28:29 +01:00
439fcaf810
feat(router): add prometheus metrics scrape endpoint ( #71 )
2023-02-16 17:18:53 +01:00
5437d49beb
feat(router): add max_total_tokens and empty_input validation ( #68 )
...
closes #65
2023-02-15 21:56:59 +01:00
9af454142a
feat: add distributed tracing ( #62 )
2023-02-13 13:02:45 +01:00
e520d5b349
fixed SSE naming ( #61 )
...
https://en.wikipedia.org/wiki/Server-sent_events
2023-02-08 22:30:11 +01:00
2fe5e1b30e
V0.2.1 ( #58 )
2023-02-07 15:40:25 +01:00
20c3c5940c
feat(router): refactor API and add openAPI schemas ( #53 )
2023-02-03 12:43:37 +01:00
b1482d9048
breaking(router): modify /generate API to only return generated text ( #50 )
...
@njhill, @yk FYI
generated_text was concatenated to the user prompt for legacy reason. We
want to remove this behaviour as we don't think it is useful and even
detrimonial to usability.
We also remove the unused Vec.
2023-02-02 15:02:04 +01:00
7b870e1e18
feat(router): use background task to manage request queue ( #52 )
...
Co-authored-by: Nick Hill <nickhill@us.ibm.com>
2023-02-02 14:59:27 +01:00
313194f6d7
feat(server): support repetition penalty ( #47 )
2023-02-01 15:58:42 +01:00
017a2a8c2f
feat: Add token streaming using ServerSideEvents support ( #41 )
2023-01-31 17:04:00 +01:00
54fec93193
fix(server): fix seeding with multiple shards ( #44 )
2023-01-31 16:01:15 +01:00
4f9ac67cfa
Revert "feat: Add token streaming using ServerSideEvents support" ( #40 )
...
Reverts huggingface/text-generation-inference#36
2023-01-31 14:21:51 +01:00
7fbfbb0dc5
feat: Add token streaming using ServerSideEvents support ( #36 )
...
Add token streaming using ServerSideEvents (SSE).
The signature of the SSE events is:
```rust
struct Details {
finish_reason: String,
generated_tokens: u32,
seed: Option<u64>,
}
struct StreamResponse {
token: Token,
generated_text: Option<String>,
details: Option<Details>,
}
struct ErrorResponse {
error: String,
}
```
2023-01-31 11:49:43 +01:00
cd298bc5e5
feat: Support sampling seeding ( #37 )
...
Co-authored-by: Yannic Kilcher <yk@users.noreply.github.com>
2023-01-30 15:36:16 +01:00
1539d3cbbe
feat(router): Remove second lock from batcher hot path ( #27 )
...
@njhill
2023-01-26 16:29:13 +01:00
5c01e2544c
fix(router): fix api-inference deployment ( #31 )
2023-01-23 17:42:14 +01:00
f9d0ec376a
feat(docker): Make the image compatible with api-inference ( #29 )
2023-01-23 17:11:27 +01:00
15511edc01
feat(server): Support SantaCoder ( #26 )
2023-01-20 12:24:39 +01:00
f7ac394935
fix(router): Obey max batch size ( #23 )
2023-01-17 09:11:21 +01:00
e6d3eb5d5d
fix(server): Minor refactorization using new_zeros ( #24 )
...
- Fix some type hints, in particular base tokenizer class
- Make use of `tensor.new_zero/empty` methods
- Simplify env var string parsing in launcher
2023-01-17 09:10:22 +01:00
60472f9d2b
feat(router): Add const parameters to validation logic ( #15 )
...
I noticed some opportunity to collapse some of the logic, in case you
are interested.
2023-01-03 10:41:22 +01:00
3efa5bbbfd
fix(router): Include special tokens when tokenizing ( #14 )
...
There's currently a discrepancy in the tokenization between the router
and python server code. The latter includes special tokens but former
does not.
This results in a token count mismatch for seq2seq models such as mt0
where the tokenizer emits an EOS token at the end.
This in turn results in some unexpected/incorrect output, in particular
when batch concatenation is involved, because the python code uses the
input length passed from the router for each row.
As far as I can tell, it is better to include this token in the encoder
`input_ids`, so I guess it's best to just adjust on the router side.
2022-12-30 19:31:44 +01:00
32a253063d
feat: Return logprobs ( #8 )
2022-12-15 17:03:56 +01:00
718096f695
feat: Support stop sequences ( #7 )
2022-12-12 18:25:22 +01:00
a2985036aa
feat(server): Add model tests ( #6 )
2022-12-08 18:49:33 +01:00
31d76e238d
fix(batching): Avoid theoretical hang in batcher loop ( #5 )
...
- Avoid theoretical hang in batcher loop
- Avoid a couple of clones in the router generate method
- Keep attention mask tensors as integers
- Remove num_heads attribute
Co-authored-by: OlivierDehaene <Olivier.dehaene@gmail.com>
2022-12-05 10:10:59 +01:00
d6d5b12e03
fix(router): Handle tokenizer errors
2022-11-14 17:15:19 +01:00
91f5f86280
fix(router): Fix HTTP status codes
2022-11-14 14:34:15 +01:00
427d7cc444
feat(server): Support AutoModelForSeq2SeqLM
2022-11-04 18:03:04 +01:00
c5665f5c8b
feat(server): Support generic AutoModelForCausalLM
2022-11-04 14:22:47 +01:00
b3b7ea0d74
feat: Use json formatter by default in docker image
2022-11-02 17:29:56 +01:00
3cf6368c77
feat(server): Support all AutoModelForCausalLM on a best effort basis
2022-10-28 19:24:00 +02:00
09674e6df9
feat(server): Support bitsandbytes
2022-10-27 14:25:29 +02:00
beb552127a
feat(client): Simplify sharded logic
2022-10-22 23:40:05 +02:00
c837893370
feat(router): Add max_waiting_tokens
2022-10-21 16:40:05 +02:00
895a341d06
fix(validation): Fix error messages
2022-10-21 10:59:15 +02:00
f16f2f5ae1
v0.1.0
2022-10-20 19:14:44 +02:00
92c1ecd008
feat: Add arguments to CLI
2022-10-17 18:27:33 +02:00
5e5d8766a2
feat: Improve error handling
2022-10-17 14:59:00 +02:00
bcb53903b8
feat: Add AML deployment
2022-10-15 20:21:50 +02:00
bf99afe916
feat: Docker image
2022-10-14 15:56:21 +02:00
39df4d9975
Use axum
2022-10-11 18:14:39 +02:00
e86ecbac63
ValidationError was not correctly handled
2022-10-11 16:53:40 +02:00
4c693e6524
Refactored gRPC interface
...
Added validation logic
2022-10-11 16:50:54 +02:00
fa9a088467
Add load testing
2022-10-11 10:36:51 +02:00
295831a481
Init
2022-10-08 12:30:12 +02:00
OlivierDehaene