Tutorials

Lightning-AI · Mar 7, 2024 · 9a4d7f7 · 9a4d7f7
1 parent e94ca9e
commit 9a4d7f7
Show file tree

Hide file tree

Showing 23 changed files with 54 additions and 57 deletions.
diff --git a/litgpt/utils.py b/litgpt/utils.py
@@ -72,7 +72,7 @@ def check_valid_checkpoint_dir(checkpoint_dir: Path) -> None:
     error_message = (
         f"--checkpoint_dir {str(checkpoint_dir.absolute())!r}{problem}."
         "\nFind download instructions at https://github.com/Lightning-AI/litgpt/blob/main/tutorials\n"
-        f"{extra}\nSee all download options by running:\n python scripts/download.py"
+        f"{extra}\nSee all download options by running:\n python litgpt/scripts/download.py"
     )
     print(error_message, file=sys.stderr)
     raise SystemExit(1)

diff --git a/notebooks/falcon-inference.ipynb b/notebooks/falcon-inference.ipynb
@@ -60,7 +60,7 @@
    "outputs": [],
    "source": [
     "# download the weights\n",
-    "!python scripts/download.py --repo_id tiiuae/falcon-7b"
+    "!python litgpt/scripts/download.py --repo_id tiiuae/falcon-7b"
    ]
   },
   {

diff --git a/tests/test_utils.py b/tests/test_utils.py
@@ -43,7 +43,7 @@ def test_check_valid_checkpoint_dir(tmp_path):
 Find download instructions at https://github.com/Lightning-AI/litgpt/blob/main/tutorials
 
 See all download options by running:
- python scripts/download.py
+ python litgpt/scripts/download.py
     """.strip()
     assert out == expected
 
@@ -57,7 +57,7 @@ def test_check_valid_checkpoint_dir(tmp_path):
 Find download instructions at https://github.com/Lightning-AI/litgpt/blob/main/tutorials
 
 See all download options by running:
- python scripts/download.py
+ python litgpt/scripts/download.py
     """.strip()
     assert out == expected
 
@@ -75,7 +75,7 @@ def test_check_valid_checkpoint_dir(tmp_path):
  --checkpoint_dir '{str(checkpoint_dir.absolute())}'
 
 See all download options by running:
- python scripts/download.py
+ python litgpt/scripts/download.py
     """.strip()
     assert out == expected
 

diff --git a/tutorials/convert_hf_checkpoint.md b/tutorials/convert_hf_checkpoint.md
@@ -1,9 +1,9 @@
 # Converting Hugging Face Transformers to LitGPT weights
 
-By default, the `scripts/download.py` script converts the downloaded HF checkpoint files into a LitGPT compatible format after downloading. For example,
+By default, the `litgpt/scripts/download.py` script converts the downloaded HF checkpoint files into a LitGPT compatible format after downloading. For example,
 
 ```bash
-python scripts/download.py --repo_id EleutherAI/pythia-14m
+python litgpt/scripts/download.py --repo_id EleutherAI/pythia-14m
 ```
 
 creates the following files:
@@ -23,12 +23,12 @@ checkpoints/
 
 
 
-To disable the automatic conversion, which is useful for development and debugging purposes, you can run the `scripts/download.py` with the `--convert_checkpoint false` flag. This will only download the checkpoint files but do not convert them for use in LitGPT:
+To disable the automatic conversion, which is useful for development and debugging purposes, you can run the `litgpt/scripts/download.py` with the `--convert_checkpoint false` flag. This will only download the checkpoint files but do not convert them for use in LitGPT:
 
 ```bash
 rm -rf checkpoints/EleutherAI/pythia-14m 
 
-python scripts/download.py \
+python litgpt/scripts/download.py \
   --repo_id EleutherAI/pythia-14m \
   --convert_checkpoint false
 
@@ -46,10 +46,10 @@ ls checkpoints/EleutherAI/pythia-14m
         └── tokenizer_config.json
 ```
 
-The required files `lit_config.json` and `lit_model.pth` files can then be manually generated via the `scripts/convert_hf_checkpoint.py` script:
+The required files `lit_config.json` and `lit_model.pth` files can then be manually generated via the `litgpt/scripts/convert_hf_checkpoint.py` script:
 
 ```bash
-python scripts/convert_hf_checkpoint.py \
+python litgpt/scripts/convert_hf_checkpoint.py \
   --checkpoint_dir checkpoints/EleutherAI/pythia-14m
 ```
 
diff --git a/tutorials/convert_lit_models.md b/tutorials/convert_lit_models.md
@@ -5,7 +5,7 @@ LitGPT weights need to be converted to a format that Hugging Face understands wi
 We provide a helpful script to convert models LitGPT models back to their equivalent Hugging Face Transformers format:
 
 ```sh
-python scripts/convert_lit_checkpoint.py \
+python litgpt/scripts/convert_lit_checkpoint.py \
     --checkpoint_path checkpoints/repo_id/lit_model.pth \
     --output_path output_path/converted.pth \
     --config_path checkpoints/repo_id/config.json
@@ -70,12 +70,12 @@ export repo_id=TinyLlama/TinyLlama-1.1B-intermediate-step-1431k-3T
 
 Instead of using TinyLlama, you can replace the `repo_id` target with any other model repository 
 specifier that is currently supported by LitGPT. You can get a list of supported repository specifier
-by running `scripts/download.py` without any additional arguments.
+by running `litgpt/scripts/download.py` without any additional arguments.
 
 Then, we download the model we specified via `$repo_id` above:
 
 ```bash
-python scripts/download.py --repo_id $repo_id
+python litgpt/scripts/download.py --repo_id $repo_id
 ```
 
 
@@ -113,7 +113,7 @@ python scripts/merge_lora.py \
 5. Convert the finetuning model back into a HF format:
 
 ```bash
-python scripts/convert_lit_checkpoint.py \
+python litgpt/scripts/convert_lit_checkpoint.py \
    --checkpoint_path $finetuned_dir/final/lit_model.pth \
    --output_path out/hf-tinyllama/converted_model.pth \
    --config_path checkpoints/$repo_id/lit_config.json 

diff --git a/tutorials/download_code_llama.md b/tutorials/download_code_llama.md
@@ -13,7 +13,7 @@ All models were  trained on 16,000 token contexts and support generations with u
 To see all the available checkpoints, run:
 
 ```bash
-python scripts/download.py | grep CodeLlama
+python litgpt/scripts/download.py | grep CodeLlama
 ```
 
 which will print
@@ -38,7 +38,7 @@ In order to use a specific checkpoint, for instance [CodeLlama-7b-Python-hf](htt
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id codellama/CodeLlama-7b-Python-hf
+python litgpt/scripts/download.py --repo_id codellama/CodeLlama-7b-Python-hf
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_dolly.md b/tutorials/download_dolly.md
@@ -12,7 +12,7 @@ For detailed info on the models, their training, and their behavior, please see
 To see all the available checkpoints for Dolly, run:
 
 ```bash
-python scripts/download.py | grep dolly
+python litgpt/scripts/download.py | grep dolly
 ```
 
 which will print
@@ -28,7 +28,7 @@ In order to use a specific Dolly checkpoint, for instance [dolly-v2-3b](https://
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id databricks/dolly-v2-3b
+python litgpt/scripts/download.py --repo_id databricks/dolly-v2-3b
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_falcon.md b/tutorials/download_falcon.md
@@ -9,7 +9,7 @@ The first Falcon release includes a base model and an instruction tuned model of
 To see all the available checkpoints for Falcon, run:
 
 ```bash
-python scripts/download.py | grep falcon
+python litgpt/scripts/download.py | grep falcon
 ```
 
 which will print
@@ -28,7 +28,7 @@ In order to use a specific Falcon checkpoint, for instance [falcon-7b](https://h
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id tiiuae/falcon-7b
+python litgpt/scripts/download.py --repo_id tiiuae/falcon-7b
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_freewilly_2.md b/tutorials/download_freewilly_2.md
@@ -7,7 +7,7 @@ FreeWilly2 leverages the Llama 2 70B foundation model to reach a performance tha
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id stabilityai/FreeWilly2
+python litgpt/scripts/download.py --repo_id stabilityai/FreeWilly2
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_function_calling_llama_2.md b/tutorials/download_function_calling_llama_2.md
@@ -10,7 +10,7 @@ In order to use the checkpoint, download the weights and convert the checkpoint
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id Trelis/Llama-2-7b-chat-hf-function-calling-v2
+python litgpt/scripts/download.py --repo_id Trelis/Llama-2-7b-chat-hf-function-calling-v2
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_gemma.md b/tutorials/download_gemma.md
@@ -8,7 +8,7 @@ For more information, please see the [technical report](https://storage.googleap
 To see all the available checkpoints, run:
 
 ```bash
-python scripts/download.py | grep gemma
+python litgpt/scripts/download.py | grep gemma
 ```
 
 which will print
@@ -30,7 +30,7 @@ After access is granted, you can find your HF hub token in <https://huggingface.
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id google/gemma-2b --access_token your_hf_token
+python litgpt/scripts/download.py --repo_id google/gemma-2b --access_token your_hf_token
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_llama_2.md b/tutorials/download_llama_2.md
@@ -13,7 +13,7 @@ Llama 2 comes in a range of parameter sizes — 7B, 13B, and 70B — as well as
 To see all the available checkpoints, run:
 
 ```bash
-python scripts/download.py | grep Llama-2
+python litgpt/scripts/download.py | grep Llama-2
 ```
 
 which will print
@@ -35,7 +35,7 @@ After access is granted, you can find your HF hub token in <https://huggingface.
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id meta-llama/Llama-2-7b-chat-hf --access_token your_hf_token
+python litgpt/scripts/download.py --repo_id meta-llama/Llama-2-7b-chat-hf --access_token your_hf_token
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_longchat.md b/tutorials/download_longchat.md
@@ -6,7 +6,7 @@ The technique used to extend the context length is described in [this blogpost](
 To see all the available checkpoints, run:
 
 ```bash
-python scripts/download.py | grep longchat
+python litgpt/scripts/download.py | grep longchat
 ```
 
 which will print
@@ -21,7 +21,7 @@ In order to use a specific checkpoint, for instance [longchat-7b-16k](https://hu
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id lmsys/longchat-7b-16k
+python litgpt/scripts/download.py --repo_id lmsys/longchat-7b-16k
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_mistral.md b/tutorials/download_mistral.md
@@ -16,7 +16,7 @@ Details about the data used to train the model or training procedure have not be
 To see all the available checkpoints, run:
 
 ```bash
-python scripts/download.py | grep -E 'Mistral|Mixtral'
+python litgpt/scripts/download.py | grep -E 'Mistral|Mixtral'
 ```
 
 which will print
@@ -34,7 +34,7 @@ In order to use the Mistral 7B model checkpoint, which requires about 14 GB of d
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id mistralai/Mistral-7B-Instruct-v0.2
+python litgpt/scripts/download.py --repo_id mistralai/Mistral-7B-Instruct-v0.2
 ```
 
 You're done! To execute the model just run:
@@ -57,7 +57,7 @@ In order to use the Mixtral 7B model checkpoint, which requires about 94 GB of d
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id mistralai/Mixtral-8x7B-Instruct-v0.1
+python litgpt/scripts/download.py --repo_id mistralai/Mixtral-8x7B-Instruct-v0.1
 ```
 
 Due to the size of the model, currently only the multi-device sequential generation script can handle it.

diff --git a/tutorials/download_openllama.md b/tutorials/download_openllama.md
@@ -7,7 +7,7 @@ The weights can serve as the drop in replacement of LLaMA in existing implementa
 To see all the available checkpoints for Open LLaMA, run:
 
 ```bash
-python scripts/download.py | grep open_llama
+python litgpt/scripts/download.py | grep open_llama
 ```
 
 which will print
@@ -23,9 +23,7 @@ In order to use a specific OpenLLaMA checkpoint, for instance [open_llama_3b](ht
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id openlm-research/open_llama_3b
-
-python scripts/convert_hf_checkpoint.py --checkpoint_dir checkpoints/openlm-research/open_llama_3b
+python litgpt/scripts/download.py --repo_id openlm-research/open_llama_3b
 ```
 
 By default, the convert_hf_checkpoint step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_phi.md b/tutorials/download_phi.md
@@ -10,7 +10,7 @@ To download the model weights and convert them to the litgpt format, run
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id microsoft/phi-2
+python litgpt/scripts/download.py --repo_id microsoft/phi-2
 ```
 
 > [!WARNING]
@@ -61,7 +61,7 @@ In order to use the phi-1.5 model checkpoint, which requires about 3 Gb of disk
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id microsoft/phi-1_5
+python litgpt/scripts/download.py --repo_id microsoft/phi-1_5
 ```
 
 You're done! To execute the model just run:

diff --git a/tutorials/download_pythia.md b/tutorials/download_pythia.md
@@ -8,7 +8,7 @@ It includes a suite of 8 checkpoints (weights) on 2 different datasets: [The Pil
 To see all the available checkpoints for Pythia, run:
 
 ```bash
-python scripts/download.py | grep pythia
+python litgpt/scripts/download.py | grep pythia
 ```
 
 which will print
@@ -39,7 +39,7 @@ In order to use a specific Pythia checkpoint, for instance [pythia-1b](https://h
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id EleutherAI/pythia-1b
+python litgpt/scripts/download.py --repo_id EleutherAI/pythia-1b
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_redpajama_incite.md b/tutorials/download_redpajama_incite.md
@@ -7,7 +7,7 @@ The release includes a base model, a chat fine-tuned model, and an instruction t
 To see all the available checkpoints for RedPajama-INCITE, run:
 
 ```bash
-python scripts/download.py | grep RedPajama
+python litgpt/scripts/download.py | grep RedPajama
 ```
 
 which will print
@@ -29,7 +29,7 @@ In order to use a specific RedPajama-INCITE checkpoint, for instance [RedPajama-
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id togethercomputer/RedPajama-INCITE-Base-3B-v1
+python litgpt/scripts/download.py --repo_id togethercomputer/RedPajama-INCITE-Base-3B-v1
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_stablecode.md b/tutorials/download_stablecode.md
@@ -11,7 +11,7 @@ For more info on the models, please visit the [StableCode repository](https://hu
 To see all the available checkpoints for StableCode, run:
 
 ```bash
-python scripts/download.py | grep -E "stable-?code"
+python litgpt/scripts/download.py | grep -E "stable-?code"
 ```
 
 which will print:
@@ -29,7 +29,7 @@ In order to use a specific StableCode checkpoint, for instance [stable-code-3b](
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
 export repo_id=stabilityai/stable-code-3b
-python scripts/download.py --repo_id $repo_id
+python litgpt/scripts/download.py --repo_id $repo_id
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM

diff --git a/tutorials/download_stablelm.md b/tutorials/download_stablelm.md
@@ -5,7 +5,7 @@ StableLM is a family of generative language models trained by StabilityAI.
 To see all the available checkpoints for StableLM, run:
 
 ```bash
-python scripts/download.py | grep stablelm
+python litgpt/scripts/download.py | grep stablelm
 ```
 
 which will print:
@@ -24,7 +24,7 @@ In order to use a specific StableLM checkpoint, for instance [stablelm-base-alph
 ```bash
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
-python scripts/download.py --repo_id stabilityai/stablelm-base-alpha-3b
+python litgpt/scripts/download.py --repo_id stabilityai/stablelm-base-alpha-3b
 ```
 
 By default, the checkpoint conversion step will use the data type of the HF checkpoint's parameters. In cases where RAM
@@ -56,8 +56,7 @@ In order to use a specific StableLM checkpoint, for instance [StableLM Zephyr 3B
 pip install 'huggingface_hub[hf_transfer] @ git+https://github.com/huggingface/huggingface_hub'
 
 export repo_id=stabilityai/stablelm-zephyr-3b
-python scripts/download.py --repo_id $repo_id
-python scripts/convert_hf_checkpoint.py --checkpoint_dir checkpoints/$repo_id
+python litgpt/scripts/download.py --repo_id $repo_id
 ```
 
 By default, the `convert_hf_checkpoint` step will use the data type of the HF checkpoint's parameters. In cases where RAM