**Title: Guide to Hugging Face and OLMo Language Models**

**Introduction:**
Hugging Face is a platform known for its open-source library providing pre-trained state-of-the-art Natural Language Processing Transformers, providing a range of functionalities for both users and developers.

**OLMo Understanding:**
OLMo is a series of Open Language Models trained on the Dolma dataset. A total of three models, namely OLMo 1B, OLMo 7B, and OLMo 7B Twin 2T, are the core models released. These models are developed to enable the science of language models.

**Model Details:**
Each model has different configurations in terms of the size, training tokens, layers, hidden size, and attention heads. The naming convention for checkpoints is step1000-tokens4B.

**Accessing Model Revisions:**
Users can easily load specific model revisions with the HuggingFace platform by adding the ‘revision’ argument.

**Inference:**
The guide provides instructions for quickly running inference with necessary installations. Users can get inference running using the provided code snippet.

**Fine-tuning:**
The guide elucidates the process of model fine-tuning from the final checkpoint and lists the available resources for further fine-tuning support.

**Evaluation:**
The guide displays the core model results for the 1B and 7B models, highlighting their performance against other models.

**Model Details Part 2:**
The guide provides additional details related to data used, model architecture, and hyperparameters as well as environmental impact, bias, risks, and limitations associated with the models.

**Citation:**
Users are encouraged to cite the model if used for their research or any other purposes. It provides BibTeX and APA citations for easy reference.

**Model Card Contact:**
For any errors in the model card, users are directed to reach out to specific contacts for assistance.

This guide gives an overview of the OLMo language models and provides users with specific details about accessing, using, and fine-tuning these models on the HuggingFace platform.

Source link
# HuggingFace OLMo Manual

OLMo is a series of Open Language Models designed to enable the science of language models. The OLMo models are trained on the Dolma dataset. We release all code, checkpoints, logs (coming soon), and details involved in training these models.

## Model Details
The core models released in this batch are the following:

– [OLMo 1B](https://huggingface.co/allenai/OLMo-1B): Trained on 3 Trillion tokens with 16 layers, 2048 hidden size, 16 attention heads, and a context length of 2048.
– [OLMo 7B](https://huggingface.co/allenai/OLMo-7B): Trained on 2.5 Trillion tokens with 32 layers, 4096 hidden size, 32 attention heads, and a context length of 2048.
– [OLMo 7B Twin 2T](https://huggingface.co/allenai/OLMo-7B-Twin-2T): Trained on 2 Trillion tokens with 32 layers, 4096 hidden size, 32 attention heads, and a context length of 2048.

We are releasing many checkpoints for these models, for every 1000 training steps. The naming convention is `step1000-tokens4B`. In particular, we focus on four revisions of the 7B models:

– OLMo 7B: The base OLMo 7B model
– OLMo 7B (not annealed): Learning rate not annealed to 0
– OLMo 7B-2T: OLMo checkpoint at 2T tokens
– OLMo-7B-Twin-2T: Twin version on different hardware

To load a specific model revision with HuggingFace, simply add the argument `revision`:

“`python
import hf_olmo
olmo = AutoModelForCausalLM.from_pretrained(“allenai/OLMo-7B”, revision=”step1000-tokens4B”)
“`

All revisions/branches can be listed in the file `revisions.txt`. Alternatively, you can access all the revisions for the models via the following code snippet:

“`python
from huggingface_hub import list_repo_refs
out = list_repo_refs(“allenai/OLMo-7B”)
branches = [b.name for b in out.branches]
“`

A few revisions were lost due to an error, but the vast majority are present.

## Model Description
– Developed by: Allen Institute for AI (AI2)
– Supported by: Databricks, Kempner Institute for the Study of Natural and Artificial Intelligence at Harvard University, AMD, CSC (Lumi Supercomputer), UW
– Model type: a Transformer style autoregressive language model.
– Language(s) (NLP): English
– License: The code and model are released under Apache 2.0
– Contact: Technical inquiries: `olmo at allenai dot org`. Press: `press at allenai dot org`
– Date cutoff: Feb./March 2023 based on Dolma dataset version.

OLMo Logo

OLMo is a series of Open Language Models designed to enable the science of language models.
The OLMo models are trained on the Dolma dataset.
We release all code, checkpoints, logs (coming soon), and details involved in training these models.



Model Details

The core models released in this batch are the following:

Size Training Tokens Layers Hidden Size Attention Heads Context Length
OLMo 1B 3 Trillion 16 2048 16 2048
OLMo 7B 2.5 Trillion 32 4096 32 2048
OLMo 7B Twin 2T 2 Trillion 32 4096 32 2048

We are releasing many checkpoints for these models, for every 1000 traing steps.
The naming convention is step1000-tokens4B.
In particular, we focus on four revisions of the 7B models:

Name HF Repo Model Revision Tokens Note
OLMo 7B allenai/OLMo-7B main 2.5T The base OLMo 7B model
OLMo 7B (not annealed) allenai/OLMo-7B step556000-tokens2460B 2.5T learning rate not annealed to 0
OLMo 7B-2T allenai/OLMo-7B step452000-tokens2000B 2T OLMo checkpoint at 2T tokens
OLMo-7B-Twin-2T allenai/OLMo-7B-Twin-2T main 2T Twin version on different hardware

To load a specific model revision with HuggingFace, simply add the argument revision:

import hf_olmo 
olmo = AutoModelForCausalLM.from_pretrained("allenai/OLMo-7B", revision="step1000-tokens4B")

All revisions/branches are listed in the file revisions.txt.
Or, you can access all the revisions for the models via the following code snippet:

from huggingface_hub import list_repo_refs
out = list_repo_refs("allenai/OLMo-7B")
branches = [b.name for b in out.branches]

A few revisions were lost due to an error, but the vast majority are present.



Model Description

  • Developed by: Allen Institute for AI (AI2)
  • Supported by: Databricks, Kempner Institute for the Study of Natural and Artificial Intelligence at Harvard University, AMD, CSC (Lumi Supercomputer), UW
  • Model type: a Transformer style autoregressive language model.
  • Language(s) (NLP): English
  • License: The code and model are released under Apache 2.0.
  • Contact: Technical inquiries: olmo at allenai dot org. Press: press at allenai dot org
  • Date cutoff: Feb./March 2023 based on Dolma dataset version.



Model Sources



Uses



Inference

Quickly get inference running with the following required installation:

pip install ai2-olmo

Now, proceed as usual with HuggingFace:

import hf_olmo

from transformers import AutoModelForCausalLM, AutoTokenizer
olmo = AutoModelForCausalLM.from_pretrained("allenai/OLMo-7B")
tokenizer = AutoTokenizer.from_pretrained("allenai/OLMo-7B")
message = ["Language modeling is "]
inputs = tokenizer(message, return_tensors='pt', return_token_type_ids=False)



response = olmo.generate(**inputs, max_new_tokens=100, do_sample=True, top_k=50, top_p=0.95)
print(tokenizer.batch_decode(response, skip_special_tokens=True)[0])
>> 'Language modeling is the first step to build natural language generation...'

Alternatively, with the pipeline abstraction:

import hf_olmo

from transformers import pipeline
olmo_pipe = pipeline("text-generation", model="allenai/OLMo-7B")
print(olmo_pipe("Language modeling is "))
>> 'Language modeling is a branch of natural language processing that aims to...'

Or, you can make this slightly faster by quantizing the model, e.g. AutoModelForCausalLM.from_pretrained("allenai/OLMo-7B", torch_dtype=torch.float16, load_in_8bit=True) (requires bitsandbytes).
The quantized model is more sensitive to typing / cuda, so it is recommended to pass the inputs as inputs.input_ids.to('cuda') to avoid potential issues.

Note, you may see the following error if ai2-olmo is not installed correctly, which is caused by internal Python check naming. We’ll update the code soon to make this error clearer.

    raise ImportError(
ImportError: This modeling file requires the following packages that were not found in your environment: hf_olmo. Run `pip install hf_olmo`



Fine-tuning

Model fine-tuning can be done from the final checkpoint (the main revision of this model) or many intermediate checkpoints. Two recipes for tuning are available.

  1. Fine-tune with the OLMo repository:
torchrun --nproc_per_node=8 scripts/train.py {path_to_train_config} \
    --data.paths=[{path_to_data}/input_ids.npy] \
    --data.label_mask_paths=[{path_to_data}/label_mask.npy] \
    --load_path={path_to_checkpoint} \
    --reset_trainer_state

For more documentation, see the GitHub readme.

  1. Further fine-tuning support is being developing in AI2’s Open Instruct repository. Details are here.



Evaluation

Core model results for the 7B model are found below.

Llama 7B Llama 2 7B Falcon 7B MPT 7B OLMo 7B (ours)
arc_challenge 44.5 39.8 47.5 46.5 48.5
arc_easy 57.0 57.7 70.4 70.5 65.4
boolq 73.1 73.5 74.6 74.2 73.4
copa 85.0 87.0 86.0 85.0 90
hellaswag 74.5 74.5 75.9 77.6 76.4
openbookqa 49.8 48.4 53.0 48.6 50.2
piqa 76.3 76.4 78.5 77.3 78.4
sciq 89.5 90.8 93.9 93.7 93.8
winogrande 68.2 67.3 68.9 69.9 67.9
Core tasks average 68.7 68.4 72.1 71.5 71.6
truthfulQA (MC2) 33.9 38.5 34.0 33 36.0
MMLU (5 shot MC) 31.5 45.0 24.0 30.8 28.3
GSM8k (mixed eval.) 10.0 (8shot CoT) 12.0 (8shot CoT) 4.0 (5 shot) 4.5 (5 shot) 8.5 (8shot CoT)
Full average 57.8 59.3 59.2 59.3 59.8

And for the 1B model:

task random StableLM 2 1.6b* Pythia 1B TinyLlama 1.1B OLMo 1B (ours)
arc_challenge 25 43.81 33.11 34.78 34.45
arc_easy 25 63.68 50.18 53.16 58.07
boolq 50 76.6 61.8 64.6 60.7
copa 50 84 72 78 79
hellaswag 25 68.2 44.7 58.7 62.5
openbookqa 25 45.8 37.8 43.6 46.4
piqa 50 74 69.1 71.1 73.7
sciq 25 94.7 86 90.5 88.1
winogrande 50 64.9 53.3 58.9 58.9
Average 36.11 68.41 56.44 61.48 62.42

*Unlike OLMo, Pythia, and TinyLlama, StabilityAI has not disclosed yet the data StableLM was trained on, making comparisons with other efforts challenging.



Model Details



Data

For training data details, please see the Dolma documentation.



Architecture

OLMo 7B architecture with peer models for comparison.

OLMo 7B Llama 2 7B OpenLM 7B Falcon 7B PaLM 8B
d_model 4096 4096 4096 4544 4096
num heads 32 32 32 71 16
num layers 32 32 32 32 32
MLP ratio ~8/3 ~8/3 ~8/3 4 4
LayerNorm type non-parametric LN RMSNorm parametric LN parametric LN parametric LN
pos embeddings RoPE RoPE RoPE RoPE RoPE
attention variant full GQA full MQA MQA
biases none none in LN only in LN only none
block type sequential sequential sequential parallel parallel
activation SwiGLU SwiGLU SwiGLU GeLU SwiGLU
sequence length 2048 4096 2048 2048 2048
batch size (instances) 2160 1024 2048 2304 512
batch size (tokens) ~4M ~4M ~4M ~4M ~1M
weight tying no no no no yes



Hyperparameters

AdamW optimizer parameters are shown below.

Size Peak LR Betas Epsilon Weight Decay
1B 4.0E-4 (0.9, 0.95) 1.0E-5 0.1
7B 3.0E-4 (0.9, 0.99) 1.0E-5 0.1

Optimizer settings comparison with peer models.

OLMo 7B Llama 2 7B OpenLM 7B Falcon 7B
warmup steps 5000 2000 2000 1000
peak LR 3.0E-04 3.0E-04 3.0E-04 6.0E-04
minimum LR 3.0E-05 3.0E-05 3.0E-05 1.2E-05
weight decay 0.1 0.1 0.1 0.1
beta1 0.9 0.9 0.9 0.99
beta2 0.95 0.95 0.95 0.999
epsilon 1.0E-05 1.0E-05 1.0E-05 1.0E-05
LR schedule linear cosine cosine cosine
gradient clipping global 1.0 global 1.0 global 1.0 global 1.0
gradient reduce dtype FP32 FP32 FP32 BF16
optimizer state dtype FP32 most likely FP32 FP32 FP32



Environmental Impact

OLMo 7B variants were either trained on MI250X GPUs at the LUMI supercomputer, or A100-40GB GPUs provided by MosaicML.
A summary of the environmental impact. Further details are available in the paper.

GPU Type Power Consumption From GPUs Carbon Intensity (kg CO₂e/KWh) Carbon Emissions (tCO₂eq)
OLMo 7B Twin MI250X (LUMI supercomputer) 135 MWh 0* 0*
OLMo 7B A100-40GB (MosaicML) 104 MWh 0.656 75.05



Bias, Risks, and Limitations

Like any base language model or fine-tuned model without safety filtering, it is relatively easy for a user to prompt these models to generate harmful and generally sensitive content.
Such content can also be produced unintentionally, especially in the case of bias, so we recommend users consider the risks of applications of this technology.

Otherwise, many facts from OLMo or any LLM will often not be true, so they should be checked.



Citation

BibTeX:

@article{Groeneveld2023OLMo,
  title={OLMo: Accelerating the Science of Language Models},
  author={Groeneveld, Dirk and Beltagy, Iz and Walsh, Pete and Bhagia, Akshita and Kinney, Rodney and Tafjord, Oyvind and Jha, Ananya Harsh and Ivison, Hamish and Magnusson, Ian and Wang, Yizhong and Arora, Shane and Atkinson, David and Authur, Russell and Chandu, Khyathi and Cohan, Arman and Dumas, Jennifer and Elazar, Yanai and Gu, Yuling and Hessel, Jack and Khot, Tushar and Merrill, William and Morrison, Jacob and Muennighoff, Niklas and Naik, Aakanksha and Nam, Crystal and Peters, Matthew E. and Pyatkin, Valentina and Ravichander, Abhilasha and Schwenk, Dustin and Shah, Saurabh and Smith, Will and Subramani, Nishant and Wortsman, Mitchell and Dasigi, Pradeep and Lambert, Nathan and Richardson, Kyle and Dodge, Jesse and Lo, Kyle and Soldaini, Luca and Smith, Noah A. and Hajishirzi, Hannaneh},
  journal={Preprint},
  year={2024}
}

APA:

Groeneveld, D., Beltagy, I., Walsh, P., Bhagia, A., Kinney, R., Tafjord, O., Jha, A., Ivison, H., Magnusson, I., Wang, Y., Arora, S., Atkinson, D., Authur, R., Chandu, K., Cohan, A., Dumas, J., Elazar, Y., Gu, Y., Hessel, J., Khot, T., Merrill, W., Morrison, J., Muennighoff, N., Naik, A., Nam, C., Peters, M., Pyatkin, V., Ravichander, A., Schwenk, D., Shah, S., Smith, W., Subramani, N., Wortsman, M., Dasigi, P., Lambert, N., Richardson, K., Dodge, J., Lo, K., Soldaini, L., Smith, N., & Hajishirzi, H. (2024). OLMo: Accelerating the Science of Language Models. Preprint.



Model Card Contact

For errors in this model card, contact Nathan or Akshita, {nathanl, akshitab} at allenai dot org.

The use cases of the

can be connected to the content included under this tag. A few potential use cases could include:

1. Displaying a Table: The

tag can be used to encapsulate a table that provides models and their related details. For example, one of the use cases might be to show the model size, training tokens, layers, hidden size, and attention heads for different language models.

2. Embedding Images: The

tag can also be utilized to contain and display an enclosed tag. This is useful for rendering images or logos related to the language models within the content.

3. Grouping Large Content: Sometimes, long or extensive content like tables, code snippets, and other information can be grouped under

tags to control the layout, size, and arrangement of these elements.

4. Displaying Code Snippets: The

can be used to contain code snippets as well. In this instance, the

could contain Python code snippets that provide examples of how to load a specific model revision using HuggingFace.

5. Presenting Model Results: The

tag is also used to display data tables, which, in this case, show the results of core model evaluations for the 7B and 1B models that are being discussed in the content.

In the mentioned HTML, the

tag is primarily used to control the structure and layout of the content that is being presented on the web page. This provides a clean and organized way to present various pieces of information from model details to environmental impact data. Ultimately, the

tag allows for better management and presentation of diverse content types within the HTML structure.

2024-02-01T18:47:19+01:00