Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Transformer models generation supports user-provided input embeddings #1276

Merged
merged 1 commit into from
Sep 25, 2024
Merged

Transformer models generation supports user-provided input embeddings #1276

merged 1 commit into from
Sep 25, 2024

Conversation

zongwave
Copy link
Contributor

@zongwave zongwave commented Aug 20, 2024 •
edited
Loading

What does this PR do?

Some multimodal generation model needs user specified embedded tokens input, for example NExT-GPT, this commit enable user-provided input embeddings for model generation.

  1. Input Embeddings Support:

Modified transformers/generation/utils.py to include logic that supports user-provided input embeddings. This allows for more flexibility in how input data can be fed into the models.

Added the --input_embeds option in example/text-generation/run_generation.py to facilitate testing with different models using embedded tokens.

  1. Testing and Performance Evaluation:

Conducted tests using the modified script on the Mistral 7B model with a batch size of 6. The tests included various input scenarios to assess the robustness and performance of the new features.

Enhanced example/text-generation/run_generation.py to support multiple decoding strategies including Greedy, Beam Search, and Contrastive Search. This allows users to specify the decoding strategy during runtime and evaluate the effectiveness of each method.

Fixes # (issue)
Optimum habana transformer models currently do not support user provided input embeddings for token generation.

Usage
To test the new features, use the following command:

python run_generation.py --model_name_or_path mistralai/Mistral-7B-Instruct-v0.2  --input_embeds --batch_size 6

Performance Data:

The table below summarizes the performance data for six popular models with both embeds and tokens input, highlighting the throughput, number of HPU graphs, graph compilation duration, and memory allocation.

<style> body { font-family: "Intel Clear", sans-serif; } table { width: 100%; border-collapse: collapse; } th, td { border: 1px solid black; padding: 8px; text-align: left; } th { background-color: #f2f2f2; } </style>
Model Name or Path Input Type Throughput (tokens/second) Number of HPU Graphs Graph Compilation Duration (seconds) Max Memory Allocated (GB)
tiiuae/falcon-7b Embeds 32.37 8 13.58 26.97
Tokens 29.77 7 13.74 26.96
mosaicml/mpt-7b Embeds 50.4 8 8.77 25.68
Tokens 50.73 7 9.39 25.68
gpt2 Embeds 93.18 8 4.68 0.65
Tokens 76.91 7 5.1 0.65
microsoft/phi-2 Embeds 31.19 8 14.62 10.48
Tokens 29.55 7 14.63 10.48
mistralai/Mistral-7B-Instruct-v0.2 Embeds 29.44 13 11.04 28.07
Tokens 25.6 11 12.09 28.07
Customized LLaMA-based Vicuna Embeds 28.28 12 10.6 25.31
Tokens 28.61 11 11.1 25.31

Before submitting

  • This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
  • Did you make sure to update the documentation with your changes?
  • [Y] Did you write any new necessary tests?

@libinta
Copy link
Collaborator

libinta commented Aug 31, 2024

@zongwave as this PR changes the common utils.py, can you run through gaudi2 ci test for text-generation?
https://github.com/huggingface/optimum-habana/blob/main/tests/test_text_generation_example.py
The way to trigger ci test
GAUDI2_CI=true RUN_SLOW=1 python -m pytest test/test_text-generation_example.py -m -k -v xxx
where xxx is specific model or test case,
please validate llama bf16/fp8, 1x/8x

@zongwave
Copy link
Contributor Author

zongwave commented Sep 5, 2024 •
edited
Loading

@zongwave as this PR changes the common utils.py, can you run through gaudi2 ci test for text-generation? https://github.com/huggingface/optimum-habana/blob/main/tests/test_text_generation_example.py The way to trigger ci test GAUDI2_CI=true RUN_SLOW=1 python -m pytest test/test_text-generation_example.py -m -k -v xxx where xxx is specific model or test case, please validate llama bf16/fp8, 1x/8x

@libinta
I ran GAUDI_CI test with command:

GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/test_text_generation_example.py -v -k meta-llama/Llama-2-7b-hf
======================================================================= test session starts =======================================================================
platform linux -- Python 3.10.12, pytest-8.3.2, pluggy-1.5.0 -- /usr/bin/python
cachedir: .pytest_cache
rootdir: /workspace/src/optimum-habana
configfile: setup.cfg
plugins: anyio-4.4.0
collected 52 items / 42 deselected / 10 selected

tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-1-True-141.25776956002076] PASSED [ 10%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-True-12808] PASSED [ 20%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-False-8711] PASSED [ 30%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-1230-False-128-128-13152.7] PASSED [ 40%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-163-False-128-2048-4774.7] PASSED [ 50%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-94-False-2048-128-1293.3] PASSED [ 60%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-81-False-2048-2048-1942.9] PASSED [ 70%]
tests/test_text_generation_example.py::test_text_generation_torch_compile[token0-meta-llama/Llama-2-7b-hf-102.27823420713148] PASSED [ 80%]
tests/test_text_generation_example.py::test_text_generation_torch_compile_distributed[token0-meta-llama/Llama-2-7b-hf-39.72973199515235] PASSED [ 90%]
tests/test_text_generation_example.py::test_text_generation_distributed_tp[token0-meta-llama/Llama-2-7b-hf-1345.2369318328463] PASSED [100%]

========================================================= 10 passed, 42 deselected in 2522.21s (0:42:02) ==========================================================

@zongwave
Copy link
Contributor Author

zongwave commented Sep 6, 2024 •
edited
Loading

@zongwave as this PR changes the common utils.py, can you run through gaudi2 ci test for text-generation? https://github.com/huggingface/optimum-habana/blob/main/tests/test_text_generation_example.py The way to trigger ci test GAUDI2_CI=true RUN_SLOW=1 python -m pytest test/test_text-generation_example.py -m -k -v xxx where xxx is specific model or test case, please validate llama bf16/fp8, 1x/8x

@libinta I ran GAUDI_CI test with command:

GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/test_text_generation_example.py -v -k meta-llama/Llama

======================================================================= test session starts =======================================================================
platform linux -- Python 3.10.12, pytest-8.3.2, pluggy-1.5.0 -- /usr/bin/python
cachedir: .pytest_cache
rootdir: /workspace/src/optimum-habana
configfile: setup.cfg
plugins: anyio-4.4.0
collected 52 items / 37 deselected / 15 selected

tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-1-True-141.25776956002076] PASSED [ 6%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-True-12808] PASSED [ 13%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-False-8711] PASSED [ 20%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-1230-False-128-128-13152.7] PASSED [ 26%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-163-False-128-2048-4774.7] PASSED [ 33%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-94-False-2048-128-1293.3] PASSED [ 40%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-81-False-2048-2048-1942.9] PASSED [ 46%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-3042-False-128-128-5374.6] PASSED [ 53%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-750-False-128-2048-7422.4] PASSED [ 60%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-207-False-2048-128-568.5] PASSED [ 66%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-8-172-False-2048-2048-4656.2] PASSED [ 73%]
tests/test_text_generation_example.py::test_text_generation_deepspeed[token0-meta-llama/Llama-2-70b-hf-8-1-64.10514998902435] PASSED [ 80%]
tests/test_text_generation_example.py::test_text_generation_torch_compile[token0-meta-llama/Llama-2-7b-hf-102.27823420713148] PASSED [ 86%]
tests/test_text_generation_example.py::test_text_generation_torch_compile_distributed[token0-meta-llama/Llama-2-7b-hf-39.72973199515235] PASSED [ 93%]
tests/test_text_generation_example.py::test_text_generation_distributed_tp[token0-meta-llama/Llama-2-7b-hf-1345.2369318328463] PASSED [100%]

========================================================= 15 passed, 37 deselected in 7759.19s (2:09:19) ==========================================================

I'm trying to figure out how to trigger llama 8x test case.

@regisss
Copy link
Collaborator

regisss commented Sep 6, 2024

I'm trying to figure out how to trigger llama 8x test case.

Llama 8x tests are this one and this one.

@zongwave
Copy link
Contributor Author

zongwave commented Sep 10, 2024 •
edited
Loading

I'm trying to figure out how to trigger llama 8x test case.

Llama 8x tests are this one and this one.

@regisss @libinta
I ran Gaudi2_CI test with command:
GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/test_text_generation_example.py -v -k llama

18 llama cases include bf16/fp8, 1x/8x configuration were selected and passed both input embeds and input tokens cases.

I triggered the CI test with "--input_embeds" option manually in test_text_generation_example.py
command += [
f"{path_to_example_dir / 'text-generation' / 'run_generation.py'}",
f"--model_name_or_path {model_name}",
f"--batch_size {batch_size}",
"--use_kv_cache",
"--input_embeds",
f"--max_new_tokens {max_output_tokens}",
]

1. Test results on input embeds generation enabled by adding option "--input_embeds" manually:

============================================================================================ test session starts =============================================================================================
platform linux -- Python 3.10.12, pytest-8.3.2, pluggy-1.5.0 -- /usr/bin/python
cachedir: .pytest_cache
rootdir: /workspace/src/optimum-habana
configfile: setup.cfg
plugins: anyio-4.4.0
collected 52 items / 34 deselected / 18 selected

tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-1-True-141.25776956002076] PASSED [ 5%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Meta-Llama-3-8B-1-True-129] PASSED [ 11%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-True-12808] PASSED [ 16%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-False-8711] PASSED [ 22%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-codellama/CodeLlama-34b-hf-1-True-32.644] PASSED [ 27%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-1230-False-128-128-13152.7] PASSED [ 33%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-163-False-128-2048-4774.7] PASSED [ 38%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-94-False-2048-128-1293.3] PASSED [ 44%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-81-False-2048-2048-1942.9] PASSED [ 50%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-3042-False-128-128-5374.6] PASSED [ 55%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-750-False-128-2048-7422.4] PASSED [ 61%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-207-False-2048-128-568.5] PASSED [ 66%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-8-172-False-2048-2048-4656.2] PASSED [ 72%]
tests/test_text_generation_example.py::test_text_generation_deepspeed[token0-meta-llama/Llama-2-70b-hf-8-1-64.10514998902435] PASSED [ 77%]
tests/test_text_generation_example.py::test_text_generation_deepspeed[token0-meta-llama/Meta-Llama-3-70B-Instruct-8-1-64] PASSED [ 83%]
tests/test_text_generation_example.py::test_text_generation_torch_compile[token0-meta-llama/Llama-2-7b-hf-102.27823420713148] PASSED [ 88%]
tests/test_text_generation_example.py::test_text_generation_torch_compile_distributed[token0-meta-llama/Llama-2-7b-hf-39.72973199515235] PASSED [ 94%]
tests/test_text_generation_example.py::test_text_generation_distributed_tp[token0-meta-llama/Llama-2-7b-hf-1345.2369318328463] PASSED [100%]
========================================================================== 18 passed, 34 deselected in 7952.60s (2:12:32) ==========================================================================a

2. Tests results on original input tokens generation:

GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/test_text_generation_example.py -v -k llama
====================================================================== test session starts =======================================================================
platform linux -- Python 3.10.12, pytest-8.3.2, pluggy-1.5.0 -- /usr/bin/python
cachedir: .pytest_cache
rootdir: /workspace/src/optimum-habana
configfile: setup.cfg
plugins: anyio-4.4.0
collected 52 items / 34 deselected / 18 selected

tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-1-True-141.25776956002076] PASSED [ 5%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Meta-Llama-3-8B-1-True-129] PASSED [ 11%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-True-12808] PASSED [ 16%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-False-8711] PASSED [ 22%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-codellama/CodeLlama-34b-hf-1-True-32.644] PASSED [ 27%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-1230-False-128-128-13152.7] PASSED [ 33%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-163-False-128-2048-4774.7] PASSED [ 38%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-94-False-2048-128-1293.3] PASSED [ 44%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-81-False-2048-2048-1942.9] PASSED [ 50%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-3042-False-128-128-5374.6] PASSED [ 55%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-750-False-128-2048-7422.4] PASSED [ 61%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-4-207-False-2048-128-568.5] PASSED [ 66%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-70b-hf-8-172-False-2048-2048-4656.2] PASSED [ 72%]
tests/test_text_generation_example.py::test_text_generation_deepspeed[token0-meta-llama/Llama-2-70b-hf-8-1-64.10514998902435] PASSED [ 77%]
tests/test_text_generation_example.py::test_text_generation_deepspeed[token0-meta-llama/Meta-Llama-3-70B-Instruct-8-1-64] PASSED [ 83%]
tests/test_text_generation_example.py::test_text_generation_torch_compile[token0-meta-llama/Llama-2-7b-hf-102.27823420713148] PASSED [ 88%]
tests/test_text_generation_example.py::test_text_generation_torch_compile_distributed[token0-meta-llama/Llama-2-7b-hf-39.72973199515235] PASSED [ 94%]
tests/test_text_generation_example.py::test_text_generation_distributed_tp[token0-meta-llama/Llama-2-7b-hf-1345.2369318328463] PASSED [100%]

========================================================= 18 passed, 34 deselected in 8202.18s (2:16:42) =========================================================

@vidyasiv
Copy link
Contributor

@regisss please take a look

@libinta libinta added the run-test Run CI for PRs from external contributors label Sep 18, 2024
@HuggingFaceDocBuilderDev
Copy link

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

regisss
regisss reviewed Sep 23, 2024
View reviewed changes
Copy link
Collaborator

@regisss regisss left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM! There is a merge conflict to solve however before we can merge.

@zongwave
Copy link
Contributor Author

LGTM! There is a merge conflict to solve however before we can merge.

@regisss @vidyasiv
I merged the code changes with #948 and verified the CI test for #948 using command:
GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/transformers/tests/models/ -k test_generate_from_inputs_embeds_decoder_only

Unfortunately, 5 test cases failed. It seems this #948 requires generating new tokens from "input_ids" and "inputs_embeds" simultaneously. FAILED tests/transformers/tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_generate_from_inputs_embeds_decoder_only - AssertionError: Lists differ: [[70,[66 chars]7, 78], [90, 71, 10, 82, 86, 98, 64, 64, 64, 6[37 chars] 64]] != [[70,[66 chars]7, 78, 39, 95, 41], [90, 71, 10, 82, 86, 98, 6[61 chars] 90]]

My PR only convers "inputs_embeds" or "input_ids" but not both at the same time. I’m curious about the real use case for this.

I need more time to make both two PRs work together.

@zongwave
Copy link
Contributor Author

zongwave commented Sep 25, 2024 •
edited
Loading

LGTM! There is a merge conflict to solve however before we can merge.

@regisss @vidyasiv I merged the code changes with #948 and verified the CI test for #948 using command: GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/transformers/tests/models/ -k test_generate_from_inputs_embeds_decoder_only

Unfortunately, 5 test cases failed. It seems this #948 requires generating new tokens from "input_ids" and "inputs_embeds" simultaneously. FAILED tests/transformers/tests/models/llama/test_modeling_llama.py::LlamaModelTest::test_generate_from_inputs_embeds_decoder_only - AssertionError: Lists differ: [[70,[66 chars]7, 78], [90, 71, 10, 82, 86, 98, 64, 64, 64, 6[37 chars] 64]] != [[70,[66 chars]7, 78, 39, 95, 41], [90, 71, 10, 82, 86, 98, 6[61 chars] 90]]

My PR only convers "inputs_embeds" or "input_ids" but not both at the same time. I’m curious about the real use case for this.

I need more time to make both two PRs work together.

@regisss @vidyasiv

I rebased the PR to the latest, in this update, transformer models support process both "input_ids" and "inputs_embeds" at the same time or separately.

Passed #948 test cases to verify the generation consistent by input tokens or embeds
root@800671-4:/workspace/src/optimum-habana# GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/transformers/tests/models/ -k test_generate_from_inputs_embeds_decoder_only
==================================================================== test session starts =====================================================================
platform linux -- Python 3.10.12, pytest-8.3.3, pluggy-1.5.0
rootdir: /workspace/src/optimum-habana
configfile: setup.cfg
plugins: anyio-4.4.0
collected 1226 items / 1218 deselected / 8 selected

tests/transformers/tests/models/bert/test_modeling_bert.py . [ 12%]
tests/transformers/tests/models/falcon/test_modeling_falcon.py . [ 25%]
tests/transformers/tests/models/gpt2/test_modeling_gpt2.py . [ 37%]
tests/transformers/tests/models/gpt_neox/test_modeling_gpt_neox.py . [ 50%]
tests/transformers/tests/models/gptj/test_modeling_gptj.py . [ 62%]
tests/transformers/tests/models/llama/test_modeling_llama.py . [ 75%]
tests/transformers/tests/models/roberta/test_modeling_roberta.py . [ 87%]
tests/transformers/tests/models/t5/test_modeling_t5.py . [ 100%]

Passed test-text-generation CI test cases for single HPU device to verify the performance reach benchmark (can not reserve 8-card Gaudi2 this time, I've ran and passed the total cases before)
root@800671-4:/workspace/src/optimum-habana# GAUDI2_CI=1 RUN_SLOW=1 python -m pytest tests/test_text_generation_example.py -v -k llama
==================================================================== test session starts =====================================================================
platform linux -- Python 3.10.12, pytest-8.3.3, pluggy-1.5.0 -- /usr/bin/python
cachedir: .pytest_cache
rootdir: /workspace/src/optimum-habana
configfile: setup.cfg
plugins: anyio-4.4.0
collected 52 items / 34 deselected / 18 selected

tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-1-True-141.25776956002076] PASSED [ 5%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Meta-Llama-3-8B-1-True-129] PASSED [ 11%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-True-12808] PASSED [ 16%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-meta-llama/Llama-2-7b-hf-512-False-8711] PASSED [ 22%]
tests/test_text_generation_example.py::test_text_generation_bf16_1x[token0-codellama/CodeLlama-34b-hf-1-True-32.644] PASSED [ 27%]

tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-1230-False-128-128-13152.7] PASSED [ 33%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-163-False-128-2048-4774.7] PASSED [ 38%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-94-False-2048-128-1293.3] PASSED [ 44%]
tests/test_text_generation_example.py::test_text_generation_fp8[token0-meta-llama/Llama-2-7b-hf-1-81-False-2048-2048-1942.9] PASSED [ 50%]

@regisss regisss merged commit d28bae9 into huggingface:main Sep 25, 2024
3 of 4 checks passed
@vidyasiv vidyasiv mentioned this pull request Sep 27, 2024
Merged
3 tasks
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@vidyasiv vidyasiv vidyasiv approved these changes

@regisss regisss regisss approved these changes

@ssarkar2 ssarkar2 Awaiting requested review from ssarkar2 ssarkar2 is a code owner

@bhargaveede bhargaveede Awaiting requested review from bhargaveede bhargaveede is a code owner

@vivekgoe vivekgoe Awaiting requested review from vivekgoe vivekgoe is a code owner

Assignees
No one assigned
Labels
run-test Run CI for PRs from external contributors
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
5 participants
@zongwave @libinta @regisss @vidyasiv @HuggingFaceDocBuilderDev