Merge branch 'main' into videollama3

Author: lkhl

.github/workflows/check_failed_tests.yml

@@ -35,7 +35,6 @@ env:
   # For gated repositories, we still need to agree to share information on the Hub repo. page in order to get access.
   # This token is created under the bot `hf-transformers-bot`.
   HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   CUDA_VISIBLE_DEVICES: 0,1
 

.github/workflows/doctest_job.yml

@@ -16,7 +16,6 @@ env:
   RUN_SLOW: yes
   OMP_NUM_THREADS: 16
   MKL_NUM_THREADS: 16
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
 
 jobs:

.github/workflows/model_jobs.yml

@@ -38,7 +38,6 @@ env:
   # For gated repositories, we still need to agree to share information on the Hub repo. page in order to get access.
   # This token is created under the bot `hf-transformers-bot`.
   HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   CUDA_VISIBLE_DEVICES: 0,1
 

.github/workflows/model_jobs_intel_gaudi.yml

@@ -26,7 +26,6 @@ env:
   TRANSFORMERS_IS_CI: yes
   PT_ENABLE_INT64_SUPPORT: 1
   HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   HF_HOME: /mnt/cache/.cache/huggingface
 
 jobs:

.github/workflows/self-comment-ci.yml

@@ -20,7 +20,6 @@ env:
   # For gated repositories, we still need to agree to share information on the Hub repo. page in order to get access.
   # This token is created under the bot `hf-transformers-bot`.
   HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   CUDA_VISIBLE_DEVICES: 0,1
 

.github/workflows/self-scheduled-intel-gaudi.yml

@@ -26,7 +26,6 @@ env:
   TRANSFORMERS_IS_CI: yes
   PT_ENABLE_INT64_SUPPORT: 1
   HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   HF_HOME: /mnt/cache/.cache/huggingface
 
 jobs:

.github/workflows/self-scheduled.yml

@@ -48,7 +48,6 @@ env:
   # For gated repositories, we still need to agree to share information on the Hub repo. page in order to get access.
   # This token is created under the bot `hf-transformers-bot`.
   HF_HUB_READ_TOKEN: ${{ secrets.HF_HUB_READ_TOKEN }}
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   CUDA_VISIBLE_DEVICES: 0,1
   NUM_SLICES: 2

.github/workflows/ssh-runner.yml

@@ -20,7 +20,6 @@ env:
   OMP_NUM_THREADS: 8
   MKL_NUM_THREADS: 8
   RUN_SLOW: yes # For gated repositories, we still need to agree to share information on the Hub repo. page in order to get access. # This token is created under the bot `hf-transformers-bot`.
-  SIGOPT_API_TOKEN: ${{ secrets.SIGOPT_API_TOKEN }}
   TF_FORCE_GPU_ALLOW_GROWTH: true
   CUDA_VISIBLE_DEVICES: 0,1
 

ISSUES.md

@@ -153,7 +153,7 @@ You are not required to read the following guidelines before opening an issue. H
     cd examples/seq2seq
     torchrun --nproc_per_node=2 ./finetune_trainer.py \
     --model_name_or_path sshleifer/distill-mbart-en-ro-12-4 --data_dir wmt_en_ro \
-    --output_dir output_dir --overwrite_output_dir \
+    --output_dir output_dir \
     --do_train --n_train 500 --num_train_epochs 1 \
     --per_device_train_batch_size 1  --freeze_embeds \
     --src_lang en_XX --tgt_lang ro_RO --task translation \

docs/source/ar/llm_tutorial.md

@@ -60,10 +60,10 @@ pip install transformers bitsandbytes>=0.39.0 -q
 أولاً، تحتاج إلى تحميل النموذج.
 
 ```py
->>> from transformers import AutoModelForCausalLM
+>>> from transformers import AutoModelForCausalLM, BitsAndBytesConfig
 
 >>> model = AutoModelForCausalLM.from_pretrained(
-...     "mistralai/Mistral-7B-v0.1", device_map="auto", load_in_4bit=True
+...     "mistralai/Mistral-7B-v0.1", device_map="auto", quantization_config=BitsAndBytesConfig(load_in_4bit=True)
 ... )
 ```
 
@@ -113,12 +113,12 @@ pip install transformers bitsandbytes>=0.39.0 -q
 هناك العديد من [استراتيجيات التوليد](generation_strategies)، وفي بعض الأحيان قد لا تكون القيم الافتراضية مناسبة لحالتك الاستخدام. إذا لم تكن الإخراج الخاصة بك متوافقة مع ما تتوقعه، فقد قمنا بإنشاء قائمة بأكثر الأخطاء الشائعة وكيفية تجنبها.
 
 ```py
->>> from transformers import AutoModelForCausalLM, AutoTokenizer
+>>> from transformers import AutoModelForCausalLM, AutoTokenizer, BitsAndBytesConfig
 
 >>> tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
 >>> tokenizer.pad_token = tokenizer.eos_token  # Most LLMs don't have a pad token by default
 >>> model = AutoModelForCausalLM.from_pretrained(
-...     "mistralai/Mistral-7B-v0.1", device_map="auto", load_in_4bit=True
+...     "mistralai/Mistral-7B-v0.1", device_map="auto", quantization_config=BitsAndBytesConfig(load_in_4bit=True)
 ... )
 ```
 
@@ -192,7 +192,7 @@ LLMs هي [معماريات فك التشفير فقط](https://huggingface.co/l
 ```python
 >>> tokenizer = AutoTokenizer.from_pretrained("HuggingFaceH4/zephyr-7b-alpha")
 >>> model = AutoModelForCausalLM.from_pretrained(
-...     "HuggingFaceH4/zephyr-7b-alpha", device_map="auto", load_in_4bit=True
+...     "HuggingFaceH4/zephyr-7b-alpha", device_map="auto", quantization_config=BitsAndBytesConfig(load_in_4bit=True)
 ... )
 >>> set_seed(0)
 >>> prompt = """How many helicopters can a human eat in one sitting? Reply as a thug."""