Fix package naming

getting-started/inference/local_inference/chat_completion/chat_completion.py

@@ -11,9 +11,9 @@
 import torch
 from transformers import AutoTokenizer
 
-from llama_recipes.inference.chat_utils import read_dialogs_from_file
-from llama_recipes.inference.model_utils import load_model, load_peft_model
-from llama_recipes.inference.safety_utils import get_safety_checker
+from llama_cookbook.inference.chat_utils import read_dialogs_from_file
+from llama_cookbook.inference.model_utils import load_model, load_peft_model
+from llama_cookbook.inference.safety_utils import get_safety_checker
 from accelerate.utils import is_xpu_available
 
 def main(

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "llama-cookbook"
-version = "0.0.4.post1"
+version = "0.0.5"
 authors = [
   { name="Hamid Shojanazeri", email="[email protected]" },
   { name="Matthias Reso", email="[email protected]" },
@@ -36,7 +36,7 @@ exclude = [
 ]
 
 [tool.hatch.build.targets.wheel]
-packages = ["src/llama_recipes"]
+packages = ["src/llama_cookbook"]
 
 [tool.hatch.metadata.hooks.requirements_txt]
 files = ["requirements.txt"]

src/README.md

@@ -14,9 +14,6 @@ Llama-Cookbook provides a pip distribution for easy install and usage in other p
 > Ensure you use the correct CUDA version (from `nvidia-smi`) when installing the PyTorch wheels. Here we are using 11.8 as `cu118`.
 > H100 GPUs work better with CUDA >12.0
 
-> > [!NOTE]
-When importing, please use the package name `llama_recipes`. We recently renamed the package to `llama-cookbook and will update the library naming in a following release`
-
 #### Install with pip
 ```
 pip install llama-cookbook

src/docs/multi_gpu.md

@@ -23,7 +23,7 @@ This runs with the `samsum_dataset` for summarization application by default.
 
 ```bash
 
-torchrun --nnodes 1 --nproc_per_node 4  recipes/quickstart/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --use_peft --peft_method lora --output_dir Path/to/save/PEFT/model
+torchrun --nnodes 1 --nproc_per_node 4  getting-started/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --use_peft --peft_method lora --output_dir Path/to/save/PEFT/model
 
 ```
 
@@ -42,7 +42,7 @@ We use `torchrun` here to spawn multiple processes for FSDP.
 Setting `use_fast_kernels` will enable using of Flash Attention or Xformer memory-efficient kernels based on the hardware being used. This would speed up the fine-tuning job. This has been enabled in `optimum` library from HuggingFace as a one-liner API, please read more [here](https://pytorch.org/blog/out-of-the-box-acceleration/).
 
 ```bash
-torchrun --nnodes 1 --nproc_per_node 4  recipes/quickstart/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --use_peft --peft_method lora --output_dir Path/to/save/PEFT/model --use_fast_kernels
+torchrun --nnodes 1 --nproc_per_node 4  getting-started/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --use_peft --peft_method lora --output_dir Path/to/save/PEFT/model --use_fast_kernels
 ```
 
 ### Fine-tuning using FSDP Only
@@ -51,7 +51,7 @@ If interested in running full parameter finetuning without making use of PEFT me
 
 ```bash
 
-torchrun --nnodes 1 --nproc_per_node 8  recipes/quickstart/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned --fsdp_config.pure_bf16 --use_fast_kernels
+torchrun --nnodes 1 --nproc_per_node 8  getting-started/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned --fsdp_config.pure_bf16 --use_fast_kernels
 
 ```
 
@@ -69,7 +69,7 @@ If you are interested in running full parameter fine-tuning on the 70B model, yo
 
 ```bash
 
-torchrun --nnodes 1 --nproc_per_node 8 recipes/quickstart/finetuning/finetuning.py --enable_fsdp --low_cpu_fsdp --fsdp_config.pure_bf16 --model_name /path_of_model_folder/70B --batch_size_training 1 --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned
+torchrun --nnodes 1 --nproc_per_node 8 getting-started/finetuning/finetuning.py --enable_fsdp --low_cpu_fsdp --fsdp_config.pure_bf16 --model_name /path_of_model_folder/70B --batch_size_training 1 --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned
 
 ```
 
@@ -79,7 +79,7 @@ Here we use a slurm script to schedule a job with slurm over multiple nodes.
 
 ```bash
 
-sbatch recipes/quickstart/finetuning/multi_node.slurm
+sbatch getting-started/finetuning/multi_node.slurm
 # Change the num nodes and GPU per nodes in the script before running.
 
 ```
@@ -102,16 +102,16 @@ To run with each of the datasets set the `dataset` flag in the command as shown
 
 ```bash
 # grammer_dataset
-torchrun --nnodes 1 --nproc_per_node 4  recipes/quickstart/finetuning/finetuning.py --enable_fsdp  --model_name /path_of_model_folder/8B --use_peft --peft_method lora --dataset grammar_dataset --save_model --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned  --fsdp_config.pure_bf16 --output_dir Path/to/save/PEFT/model
+torchrun --nnodes 1 --nproc_per_node 4  getting-started/finetuning/finetuning.py --enable_fsdp  --model_name /path_of_model_folder/8B --use_peft --peft_method lora --dataset grammar_dataset --save_model --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned  --fsdp_config.pure_bf16 --output_dir Path/to/save/PEFT/model
 
 # alpaca_dataset
 
-torchrun --nnodes 1 --nproc_per_node 4  recipes/quickstart/finetuning/finetuning.py --enable_fsdp  --model_name /path_of_model_folder/8B --use_peft --peft_method lora --dataset alpaca_dataset --save_model --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned --fsdp_config.pure_bf16 --output_dir Path/to/save/PEFT/model
+torchrun --nnodes 1 --nproc_per_node 4  getting-started/finetuning/finetuning.py --enable_fsdp  --model_name /path_of_model_folder/8B --use_peft --peft_method lora --dataset alpaca_dataset --save_model --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned --fsdp_config.pure_bf16 --output_dir Path/to/save/PEFT/model
 
 
 # samsum_dataset
 
-torchrun --nnodes 1 --nproc_per_node 4  recipes/quickstart/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --use_peft --peft_method lora --dataset samsum_dataset --save_model --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned --fsdp_config.pure_bf16 --output_dir Path/to/save/PEFT/model
+torchrun --nnodes 1 --nproc_per_node 4  getting-started/finetuning/finetuning.py --enable_fsdp --model_name /path_of_model_folder/8B --use_peft --peft_method lora --dataset samsum_dataset --save_model --dist_checkpoint_root_folder model_checkpoints --dist_checkpoint_folder fine-tuned --fsdp_config.pure_bf16 --output_dir Path/to/save/PEFT/model
 
 ```
 
@@ -166,11 +166,11 @@ It lets us specify the training settings for everything from `model_name` to `da
     profiler_dir: str = "PATH/to/save/profiler/results" # will be used if using profiler
 ```
 
-* [Datasets config file](../llama_recipes/configs/datasets.py) provides the available options for datasets.
+* [Datasets config file](../llama_cookbook/configs/datasets.py) provides the available options for datasets.
 
-* [peft config file](../llama_recipes/configs/peft.py) provides the supported PEFT methods and respective settings that can be modified.
+* [peft config file](../llama_cookbook/configs/peft.py) provides the supported PEFT methods and respective settings that can be modified.
 
-* [FSDP config file](../llama_recipes/configs/fsdp.py) provides FSDP settings such as:
+* [FSDP config file](../llama_cookbook/configs/fsdp.py) provides FSDP settings such as:
 
     * `mixed_precision` boolean flag to specify using mixed precision, defatults to true.
 

src/docs/single_gpu.md

@@ -21,7 +21,7 @@ Get access to a machine with one GPU or if using a multi-GPU machine please make
 
 ```bash
 
-python -m llama_recipes.finetuning  --use_peft --peft_method lora --quantization 8bit --use_fp16 --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
+python -m llama_cookbook.finetuning  --use_peft --peft_method lora --quantization 8bit --use_fp16 --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
 
 ```
 The args used in the command above are:
@@ -42,7 +42,7 @@ Currently 4 datasets are supported that can be found in [Datasets config file](.
 * `alpaca_dataset` : to get this open source data please download the `aplaca.json` to `ft_dataset` folder.
 
 ```bash
-wget -P src/llama_recipes/datasets https://raw.githubusercontent.com/tatsu-lab/stanford_alpaca/main/alpaca_data.json
+wget -P src/llama_cookbook/datasets https://raw.githubusercontent.com/tatsu-lab/stanford_alpaca/main/alpaca_data.json
 ```
 
 * `samsum_dataset`
@@ -52,16 +52,16 @@ to run with each of the datasets set the `dataset` flag in the command as shown
 ```bash
 # grammer_dataset
 
-python -m llama_recipes.finetuning  --use_peft --peft_method lora --quantization 8bit --dataset grammar_dataset --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
+python -m llama_cookbook.finetuning  --use_peft --peft_method lora --quantization 8bit --dataset grammar_dataset --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
 
 # alpaca_dataset
 
-python -m llama_recipes.finetuning  --use_peft --peft_method lora --quantization 8bit --dataset alpaca_dataset --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
+python -m llama_cookbook.finetuning  --use_peft --peft_method lora --quantization 8bit --dataset alpaca_dataset --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
 
 
 # samsum_dataset
 
-python -m llama_recipes.finetuning  --use_peft --peft_method lora --quantization 8bit --dataset samsum_dataset --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
+python -m llama_cookbook.finetuning  --use_peft --peft_method lora --quantization 8bit --dataset samsum_dataset --model_name /path_of_model_folder/8B --output_dir Path/to/save/PEFT/model
 
 ```
 

src/llama_cookbook/configs/__init__.py

@@ -0,0 +1,8 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# This software may be used and distributed according to the terms of the Llama 2 Community License Agreement.
+
+from llama_cookbook.configs.peft import lora_config, llama_adapter_config, prefix_config
+from llama_cookbook.configs.fsdp import fsdp_config
+from llama_cookbook.configs.training import train_config
+from llama_cookbook.configs.wandb import wandb_config
+from llama_cookbook.configs.quantization import quantization_config

src/llama_recipes/datasets/__init__.py → src/llama_cookbook/datasets/__init__.py

@@ -3,11 +3,11 @@
 
 from functools import partial
 
-from llama_recipes.datasets.grammar_dataset.grammar_dataset import get_dataset as get_grammar_dataset
-from llama_recipes.datasets.alpaca_dataset import InstructionDataset as get_alpaca_dataset
-from llama_recipes.datasets.custom_dataset import get_custom_dataset,get_data_collator
-from llama_recipes.datasets.samsum_dataset import get_preprocessed_samsum as get_samsum_dataset
-from llama_recipes.datasets.toxicchat_dataset import get_llamaguard_toxicchat_dataset as get_llamaguard_toxicchat_dataset
+from llama_cookbook.datasets.grammar_dataset.grammar_dataset import get_dataset as get_grammar_dataset
+from llama_cookbook.datasets.alpaca_dataset import InstructionDataset as get_alpaca_dataset
+from llama_cookbook.datasets.custom_dataset import get_custom_dataset,get_data_collator
+from llama_cookbook.datasets.samsum_dataset import get_preprocessed_samsum as get_samsum_dataset
+from llama_cookbook.datasets.toxicchat_dataset import get_llamaguard_toxicchat_dataset as get_llamaguard_toxicchat_dataset
 DATASET_PREPROC = {
     "alpaca_dataset": partial(get_alpaca_dataset),
     "grammar_dataset": get_grammar_dataset,

src/llama_recipes/datasets/toxicchat_dataset.py → src/llama_cookbook/datasets/toxicchat_dataset.py

@@ -6,7 +6,7 @@
 import copy
 import datasets
 import itertools
-from llama_recipes.inference.prompt_format_utils import  LLAMA_GUARD_3_CATEGORY
+from llama_cookbook.inference.prompt_format_utils import  LLAMA_GUARD_3_CATEGORY
 import ast
 import fire
 
@@ -19,7 +19,7 @@ def tokenize_prompt_and_labels(full_prompt, tokenizer):
         return dict(combined_tokens, attention_mask=[1]*len(combined_tokens["input_ids"]))
 
 
-from llama_recipes.data.llama_guard.finetuning_data_formatter import TrainingExample, Guidelines, Category, LlamaGuardPromptConfigs, LlamaGuardGenerationConfigs, ExplanationPosition, AugmentationConfigs, FormatterConfigs, create_formatted_finetuning_examples
+from llama_cookbook.data.llama_guard.finetuning_data_formatter import TrainingExample, Guidelines, Category, LlamaGuardPromptConfigs, LlamaGuardGenerationConfigs, ExplanationPosition, AugmentationConfigs, FormatterConfigs, create_formatted_finetuning_examples
 from datasets import Dataset, DatasetInfo
 
 def mapTcCategoriesToLGCategories(TcCategoriesString):

src/llama_recipes/finetuning.py → src/llama_cookbook/finetuning.py

@@ -13,29 +13,29 @@
 import torch.optim as optim
 from accelerate.utils import is_xpu_available
 
-from llama_recipes.configs import (
+from llama_cookbook.configs import (
     fsdp_config as FSDP_CONFIG,
     quantization_config as QUANTIZATION_CONFIG,
     train_config as TRAIN_CONFIG,
 )
-from llama_recipes.data.concatenator import ConcatDataset
-from llama_recipes.policies import AnyPrecisionAdamW, apply_fsdp_checkpointing
+from llama_cookbook.data.concatenator import ConcatDataset
+from llama_cookbook.policies import AnyPrecisionAdamW, apply_fsdp_checkpointing
 
-from llama_recipes.utils import fsdp_auto_wrap_policy
-from llama_recipes.utils.config_utils import (
+from llama_cookbook.utils import fsdp_auto_wrap_policy
+from llama_cookbook.utils.config_utils import (
     check_fsdp_config,
     generate_dataset_config,
     generate_peft_config,
     get_dataloader_kwargs,
     update_config,
 )
-from llama_recipes.utils.dataset_utils import (
+from llama_cookbook.utils.dataset_utils import (
     get_custom_data_collator,
     get_preprocessed_dataset,
 )
 
-from llama_recipes.utils.fsdp_utils import hsdp_device_mesh
-from llama_recipes.utils.train_utils import (
+from llama_cookbook.utils.fsdp_utils import hsdp_device_mesh
+from llama_cookbook.utils.train_utils import (
     clear_gpu_cache,
     freeze_transformer_layers,
     freeze_LLM_only,

src/llama_recipes/inference/checkpoint_converter_fsdp_hf.py → src/llama_cookbook/inference/checkpoint_converter_fsdp_hf.py

@@ -9,7 +9,7 @@
 import fire
 import yaml
 
-from llama_recipes.inference.model_utils import load_llama_from_config
+from llama_cookbook.inference.model_utils import load_llama_from_config
 
 from transformers import AutoConfig, AutoTokenizer, MllamaProcessor
 

src/llama_recipes/inference/model_utils.py → src/llama_cookbook/inference/model_utils.py

@@ -3,8 +3,8 @@
 
 from warnings import warn
 
-from llama_recipes.configs import quantization_config as QUANT_CONFIG
-from llama_recipes.utils.config_utils import update_config
+from llama_cookbook.configs import quantization_config as QUANT_CONFIG
+from llama_cookbook.utils.config_utils import update_config
 from peft import PeftModel
 from transformers import (
     AutoConfig,

src/llama_recipes/model_checkpointing/__init__.py → src/llama_cookbook/model_checkpointing/__init__.py

@@ -1,7 +1,7 @@
 # Copyright (c) Meta Platforms, Inc. and affiliates.
 # This software may be used and distributed according to the terms of the Llama 2 Community License Agreement.
 
-from llama_recipes.model_checkpointing.checkpoint_handler import (
+from llama_cookbook.model_checkpointing.checkpoint_handler import (
     load_model_checkpoint,
     save_fsdp_model_checkpoint_full,
     save_peft_checkpoint,

src/llama_cookbook/policies/__init__.py

@@ -0,0 +1,7 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# This software may be used and distributed according to the terms of the Llama 2 Community License Agreement.
+
+from llama_cookbook.policies.mixed_precision import *
+from llama_cookbook.policies.wrapping import *
+from llama_cookbook.policies.activation_checkpointing_functions import apply_fsdp_checkpointing
+from llama_cookbook.policies.anyprecision_optimizer import AnyPrecisionAdamW

src/llama_cookbook/utils/__init__.py

@@ -0,0 +1,7 @@
+# Copyright (c) Meta Platforms, Inc. and affiliates.
+# This software may be used and distributed according to the terms of the Llama 2 Community License Agreement.
+
+from llama_cookbook.utils.memory_utils import MemoryTrace
+from llama_cookbook.utils.dataset_utils import *
+from llama_cookbook.utils.fsdp_utils import fsdp_auto_wrap_policy, hsdp_device_mesh
+from llama_cookbook.utils.train_utils import *

src/llama_recipes/utils/config_utils.py → src/llama_cookbook/utils/config_utils.py

@@ -15,9 +15,9 @@
 from transformers import default_data_collator
 from transformers.data import DataCollatorForSeq2Seq
 
-from llama_recipes.configs import datasets, lora_config, llama_adapter_config, prefix_config, train_config
-from llama_recipes.data.sampler import LengthBasedBatchSampler, DistributedLengthBasedBatchSampler
-from llama_recipes.datasets import DATASET_PREPROC
+from llama_cookbook.configs import datasets, lora_config, llama_adapter_config, prefix_config, train_config
+from llama_cookbook.data.sampler import LengthBasedBatchSampler, DistributedLengthBasedBatchSampler
+from llama_cookbook.datasets import DATASET_PREPROC
 
 def update_config(config, **kwargs):
     if isinstance(config, (tuple, list)):

src/llama_recipes/utils/dataset_utils.py → src/llama_cookbook/utils/dataset_utils.py

@@ -3,9 +3,9 @@
 
 import torch
 
-from llama_recipes.data.concatenator import ConcatDataset
-from llama_recipes.datasets import DATASET_PREPROC, DATALOADER_COLLATE_FUNC
-from llama_recipes.utils.config_utils import get_dataloader_kwargs
+from llama_cookbook.data.concatenator import ConcatDataset
+from llama_cookbook.datasets import DATASET_PREPROC, DATALOADER_COLLATE_FUNC
+from llama_cookbook.utils.config_utils import get_dataloader_kwargs
 
 
 def get_preprocessed_dataset(

src/llama_recipes/utils/train_utils.py → src/llama_cookbook/utils/train_utils.py

@@ -20,11 +20,11 @@
 import json
 
 
-from llama_recipes.model_checkpointing import save_fsdp_model_checkpoint_full, save_model_and_optimizer_sharded, save_optimizer_checkpoint, save_peft_checkpoint, save_model_checkpoint
-from llama_recipes.policies import fpSixteen,bfSixteen, get_llama_wrapper
-from llama_recipes.utils.memory_utils import MemoryTrace
+from llama_cookbook.model_checkpointing import save_fsdp_model_checkpoint_full, save_model_and_optimizer_sharded, save_optimizer_checkpoint, save_peft_checkpoint, save_model_checkpoint
+from llama_cookbook.policies import fpSixteen,bfSixteen, get_llama_wrapper
+from llama_cookbook.utils.memory_utils import MemoryTrace
 from accelerate.utils import is_xpu_available, is_ccl_available
-from llama_recipes.utils.flop_utils import FlopMeasure
+from llama_cookbook.utils.flop_utils import FlopMeasure
 def set_tokenizer_params(tokenizer: LlamaTokenizer):
     tokenizer.pad_token_id = 0
     tokenizer.padding_side = "left"