Improve the defaults for the launcher (huggingface#1727)

# What does this PR do?

- Renamed `max_input_length` into `max_input_tokens` for consistency
(backward compatible change, will yell if both are set.)
- Will now use the config for `max_input_tokens` `max_total_token` and
`max_batch_total_tokens`.
- Capping the values to 16k in order to save VRAM on behalf of users
(overriddable by simply setting the values).

<!--
Congratulations! You've made it this far! You're not quite done yet
though.

Once merged, your PR is going to appear in the release notes with the
title you set, so make sure it's a great title that fully reflects the
extent of your awesome contribution.

Then, please replace this with a description of the change and which
issue is fixed (if applicable). Please also include relevant motivation
and context. List any dependencies (if any) that are required for this
change.

Once you're done, someone will review your PR shortly (see the section
"Who can review?" below to tag some potential reviewers). They may
suggest changes to make the code even better. If no one reviewed your PR
after a week has passed, don't hesitate to post a new comment
@-mentioning the same persons---sometimes notifications get lost.
-->

<!-- Remove if not applicable -->

Fixes # (issue)


## Before submitting
- [ ] This PR fixes a typo or improves the docs (you can dismiss the
other checks if that's the case).
- [ ] Did you read the [contributor
guideline](https://github.com/huggingface/transformers/blob/main/CONTRIBUTING.md#start-contributing-pull-requests),
      Pull Request section?
- [ ] Was this discussed/approved via a Github issue or the
[forum](https://discuss.huggingface.co/)? Please add a link
      to it if that's the case.
- [ ] Did you make sure to update the documentation with your changes?
Here are the
[documentation
guidelines](https://github.com/huggingface/transformers/tree/main/docs),
and
[here are tips on formatting
docstrings](https://github.com/huggingface/transformers/tree/main/docs#writing-source-documentation).
- [ ] Did you write any new necessary tests?


## Who can review?

Anyone in the community is free to review the PR once the tests have
passed. Feel free to tag
members/contributors who may be interested in your PR.

<!-- Your PR will be replied to more quickly if you can figure out the
right person to tag with @


@OlivierDehaene OR @Narsil

 -->

docs/source/basic_tutorials/launcher.md

@@ -60,9 +60,9 @@ Options:
           [env: QUANTIZE=]
 
           Possible values:
-          - awq:              4 bit quantization. Requires a specific AWQ quantized model: https://hf.co/models?search=awq. Should replace GPTQ models wherever possible because of the better latency
-          - eetq:             8 bit quantization, doesn't require specific model. Should be a drop-in replacement to bitsandbytes with much better performance. Kernels are from https://github.com/NetEase-FuXi/EETQ.git
-          - gptq:             4 bit quantization. Requires a specific GTPQ quantized model: https://hf.co/models?search=gptq. text-generation-inference will use exllama (faster) kernels wherever possible, and use triton kernel (wider support) when it's not. AWQ has faster kernels
+          - awq:              4 bit quantization. Requires a specific AWQ quantized model: <https://hf.co/models?search=awq>. Should replace GPTQ models wherever possible because of the better latency
+          - eetq:             8 bit quantization, doesn't require specific model. Should be a drop-in replacement to bitsandbytes with much better performance. Kernels are from <https://github.com/NetEase-FuXi/EETQ.git>
+          - gptq:             4 bit quantization. Requires a specific GTPQ quantized model: <https://hf.co/models?search=gptq>. text-generation-inference will use exllama (faster) kernels wherever possible, and use triton kernel (wider support) when it's not. AWQ has faster kernels
           - bitsandbytes:     Bitsandbytes 8bit. Can be applied on any model, will cut the memory requirement in half, but it is known that the model will be much slower to run than the native f16
           - bitsandbytes-nf4: Bitsandbytes 4bit. Can be applied on any model, will cut the memory requirement by 4x, but it is known that the model will be much slower to run than the native f16
           - bitsandbytes-fp4: Bitsandbytes 4bit. nf4 should be preferred in most cases but maybe this one has better perplexity performance for you model
@@ -129,23 +129,29 @@ Options:
           [env: MAX_TOP_N_TOKENS=]
           [default: 5]
 
+```
+## MAX_INPUT_TOKENS
+```shell
+      --max-input-tokens <MAX_INPUT_TOKENS>
+          This is the maximum allowed input length (expressed in number of tokens) for users. The larger this value, the longer prompt users can send which can impact the overall memory required to handle the load. Please note that some models have a finite range of sequence they can handle. Default to min(max_position_embeddings - 1, 4095)
+          
+          [env: MAX_INPUT_TOKENS=]
+
 ```
 ## MAX_INPUT_LENGTH
 ```shell
       --max-input-length <MAX_INPUT_LENGTH>
-          This is the maximum allowed input length (expressed in number of tokens) for users. The larger this value, the longer prompt users can send which can impact the overall memory required to handle the load. Please note that some models have a finite range of sequence they can handle
+          Legacy version of [`Args::max_input_tokens`]
           
           [env: MAX_INPUT_LENGTH=]
-          [default: 1024]
 
 ```
 ## MAX_TOTAL_TOKENS
 ```shell
       --max-total-tokens <MAX_TOTAL_TOKENS>
-          This is the most important value to set as it defines the "memory budget" of running clients requests. Clients will send input sequences and ask to generate `max_new_tokens` on top. with a value of `1512` users can send either a prompt of `1000` and ask for `512` new tokens, or send a prompt of `1` and ask for `1511` max_new_tokens. The larger this value, the larger amount each request will be in your RAM and the less effective batching can be
+          This is the most important value to set as it defines the "memory budget" of running clients requests. Clients will send input sequences and ask to generate `max_new_tokens` on top. with a value of `1512` users can send either a prompt of `1000` and ask for `512` new tokens, or send a prompt of `1` and ask for `1511` max_new_tokens. The larger this value, the larger amount each request will be in your RAM and the less effective batching can be. Default to min(max_position_embeddings, 4096)
           
           [env: MAX_TOTAL_TOKENS=]
-          [default: 2048]
 
 ```
 ## WAITING_SERVED_RATIO
@@ -162,10 +168,9 @@ Options:
 ## MAX_BATCH_PREFILL_TOKENS
 ```shell
       --max-batch-prefill-tokens <MAX_BATCH_PREFILL_TOKENS>
-          Limits the number of tokens for the prefill operation. Since this operation take the most memory and is compute bound, it is interesting to limit the number of requests that can be sent
+          Limits the number of tokens for the prefill operation. Since this operation take the most memory and is compute bound, it is interesting to limit the number of requests that can be sent. Default to `max_input_tokens + 50` to give a bit of room
           
           [env: MAX_BATCH_PREFILL_TOKENS=]
-          [default: 4096]
 
 ```
 ## MAX_BATCH_TOTAL_TOKENS
@@ -210,10 +215,9 @@ Options:
 ## CUDA_GRAPHS
 ```shell
       --cuda-graphs <CUDA_GRAPHS>
-          Specify the batch sizes to compute cuda graphs for. Use "0" to disable
+          Specify the batch sizes to compute cuda graphs for. Use "0" to disable. Default = "1,2,4,8,16,32"
           
           [env: CUDA_GRAPHS=]
-          [default: 1,2,4,8,16,32,64,96,128]
 
 ```
 ## HOSTNAME

integration-tests/models/test_t5_sharded.py

@@ -3,7 +3,7 @@
 
 @pytest.fixture(scope="module")
 def t5_sharded_handle(launcher):
-    with launcher("google/flan-t5-xxl", num_shard=2) as handle:
+    with launcher("google/flan-t5-xxl", num_shard=4) as handle:
         yield handle
 
 

launcher/Cargo.toml

@@ -9,8 +9,10 @@ homepage.workspace = true
 [dependencies]
 clap = { version = "4.4.5", features = ["derive", "env"] }
 ctrlc = { version = "3.4.1", features = ["termination"] }
+hf-hub = "0.3.2"
 nix = { version = "0.28.0", features = ["signal"] }
-serde = { version = "1.0.188", features = ["derive"]  }
+once_cell = "1.19.0"
+serde = { version = "1.0.188", features = ["derive"] }
 serde_json = "1.0.107"
 tracing = "0.1.37"
 tracing-subscriber = { version = "0.3.17", features = ["json", "env-filter"] }