Datasets:

librarian-bots

/

model_cards_with_metadata_test

like 1

timm/dm_nfnet_f0.dm_in1k

timm

timm

image-classification

--- license: apache-2.0 library_name: timm tags: - image-classification - timm datasets: - imagenet-1k --- # Model card for dm_nfnet_f0.dm_in1k A NFNet (Normalization Free Network) image classification model. Trained on ImageNet-1k by paper authors. Normalization Free Networks are (pre-activation) ResNet-like models without any normalization layers. Instead of Batch Normalization or alternatives, they use Scaled Weight Standardization and specifically placed scalar gains in residual path and at non-linearities based on signal propagation analysis. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 71.5 - GMACs: 7.2 - Activations (M): 10.2 - Image size: train = 192 x 192, test = 256 x 256 - **Papers:** - High-Performance Large-Scale Image Recognition Without Normalization: https://arxiv.org/abs/2102.06171 - Characterizing signal propagation to close the performance gap in unnormalized ResNets: https://arxiv.org/abs/2101.08692 - **Original:** https://github.com/deepmind/deepmind-research/tree/master/nfnets - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('dm_nfnet_f0.dm_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'dm_nfnet_f0.dm_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 96, 96]) # torch.Size([1, 256, 48, 48]) # torch.Size([1, 512, 24, 24]) # torch.Size([1, 1536, 12, 12]) # torch.Size([1, 3072, 6, 6]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'dm_nfnet_f0.dm_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 3072, 6, 6) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{brock2021high, author={Andrew Brock and Soham De and Samuel L. Smith and Karen Simonyan}, title={High-Performance Large-Scale Image Recognition Without Normalization}, journal={arXiv preprint arXiv:2102.06171}, year={2021} } ``` ```bibtex @inproceedings{brock2021characterizing, author={Andrew Brock and Soham De and Samuel L. Smith}, title={Characterizing signal propagation to close the performance gap in unnormalized ResNets}, booktitle={9th International Conference on Learning Representations, {ICLR}}, year={2021} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

meta-llama/Meta-Llama-3-70B

meta-llama

transformers

text-generation

--- language: - en pipeline_tag: text-generation tags: - facebook - meta - pytorch - llama - llama-3 license: llama3 extra_gated_prompt: >- ### META LLAMA 3 COMMUNITY LICENSE AGREEMENT Meta Llama 3 Version Release Date: April 18, 2024 "Agreement" means the terms and conditions for use, reproduction, distribution and modification of the Llama Materials set forth herein. "Documentation" means the specifications, manuals and documentation accompanying Meta Llama 3 distributed by Meta at https://llama.meta.com/get-started/. "Licensee" or "you" means you, or your employer or any other person or entity (if you are entering into this Agreement on such person or entity’s behalf), of the age required under applicable laws, rules or regulations to provide legal consent and that has legal authority to bind your employer or such other person or entity if you are entering in this Agreement on their behalf. "Meta Llama 3" means the foundational large language models and software and algorithms, including machine-learning model code, trained model weights, inference-enabling code, training-enabling code, fine-tuning enabling code and other elements of the foregoing distributed by Meta at https://llama.meta.com/llama-downloads. "Llama Materials" means, collectively, Meta’s proprietary Meta Llama 3 and Documentation (and any portion thereof) made available under this Agreement. "Meta" or "we" means Meta Platforms Ireland Limited (if you are located in or, if you are an entity, your principal place of business is in the EEA or Switzerland) and Meta Platforms, Inc. (if you are located outside of the EEA or Switzerland). 1. License Rights and Redistribution. a. Grant of Rights. You are granted a non-exclusive, worldwide, non-transferable and royalty-free limited license under Meta’s intellectual property or other rights owned by Meta embodied in the Llama Materials to use, reproduce, distribute, copy, create derivative works of, and make modifications to the Llama Materials. b. Redistribution and Use. i. If you distribute or make available the Llama Materials (or any derivative works thereof), or a product or service that uses any of them, including another AI model, you shall (A) provide a copy of this Agreement with any such Llama Materials; and (B) prominently display “Built with Meta Llama 3” on a related website, user interface, blogpost, about page, or product documentation. If you use the Llama Materials to create, train, fine tune, or otherwise improve an AI model, which is distributed or made available, you shall also include “Llama 3” at the beginning of any such AI model name. ii. If you receive Llama Materials, or any derivative works thereof, from a Licensee as part of an integrated end user product, then Section 2 of this Agreement will not apply to you. iii. You must retain in all copies of the Llama Materials that you distribute the following attribution notice within a “Notice” text file distributed as a part of such copies: “Meta Llama 3 is licensed under the Meta Llama 3 Community License, Copyright © Meta Platforms, Inc. All Rights Reserved.” iv. Your use of the Llama Materials must comply with applicable laws and regulations (including trade compliance laws and regulations) and adhere to the Acceptable Use Policy for the Llama Materials (available at https://llama.meta.com/llama3/use-policy), which is hereby incorporated by reference into this Agreement. v. You will not use the Llama Materials or any output or results of the Llama Materials to improve any other large language model (excluding Meta Llama 3 or derivative works thereof). 2. Additional Commercial Terms. If, on the Meta Llama 3 version release date, the monthly active users of the products or services made available by or for Licensee, or Licensee’s affiliates, is greater than 700 million monthly active users in the preceding calendar month, you must request a license from Meta, which Meta may grant to you in its sole discretion, and you are not authorized to exercise any of the rights under this Agreement unless or until Meta otherwise expressly grants you such rights. 3. Disclaimer of Warranty. UNLESS REQUIRED BY APPLICABLE LAW, THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS THEREFROM ARE PROVIDED ON AN “AS IS” BASIS, WITHOUT WARRANTIES OF ANY KIND, AND META DISCLAIMS ALL WARRANTIES OF ANY KIND, BOTH EXPRESS AND IMPLIED, INCLUDING, WITHOUT LIMITATION, ANY WARRANTIES OF TITLE, NON-INFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE. YOU ARE SOLELY RESPONSIBLE FOR DETERMINING THE APPROPRIATENESS OF USING OR REDISTRIBUTING THE LLAMA MATERIALS AND ASSUME ANY RISKS ASSOCIATED WITH YOUR USE OF THE LLAMA MATERIALS AND ANY OUTPUT AND RESULTS. 4. Limitation of Liability. IN NO EVENT WILL META OR ITS AFFILIATES BE LIABLE UNDER ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, TORT, NEGLIGENCE, PRODUCTS LIABILITY, OR OTHERWISE, ARISING OUT OF THIS AGREEMENT, FOR ANY LOST PROFITS OR ANY INDIRECT, SPECIAL, CONSEQUENTIAL, INCIDENTAL, EXEMPLARY OR PUNITIVE DAMAGES, EVEN IF META OR ITS AFFILIATES HAVE BEEN ADVISED OF THE POSSIBILITY OF ANY OF THE FOREGOING. 5. Intellectual Property. a. No trademark licenses are granted under this Agreement, and in connection with the Llama Materials, neither Meta nor Licensee may use any name or mark owned by or associated with the other or any of its affiliates, except as required for reasonable and customary use in describing and redistributing the Llama Materials or as set forth in this Section 5(a). Meta hereby grants you a license to use “Llama 3” (the “Mark”) solely as required to comply with the last sentence of Section 1.b.i. You will comply with Meta’s brand guidelines (currently accessible at https://about.meta.com/brand/resources/meta/company-brand/ ). All goodwill arising out of your use of the Mark will inure to the benefit of Meta. b. Subject to Meta’s ownership of Llama Materials and derivatives made by or for Meta, with respect to any derivative works and modifications of the Llama Materials that are made by you, as between you and Meta, you are and will be the owner of such derivative works and modifications. c. If you institute litigation or other proceedings against Meta or any entity (including a cross-claim or counterclaim in a lawsuit) alleging that the Llama Materials or Meta Llama 3 outputs or results, or any portion of any of the foregoing, constitutes infringement of intellectual property or other rights owned or licensable by you, then any licenses granted to you under this Agreement shall terminate as of the date such litigation or claim is filed or instituted. You will indemnify and hold harmless Meta from and against any claim by any third party arising out of or related to your use or distribution of the Llama Materials. 6. Term and Termination. The term of this Agreement will commence upon your acceptance of this Agreement or access to the Llama Materials and will continue in full force and effect until terminated in accordance with the terms and conditions herein. Meta may terminate this Agreement if you are in breach of any term or condition of this Agreement. Upon termination of this Agreement, you shall delete and cease use of the Llama Materials. Sections 3, 4 and 7 shall survive the termination of this Agreement. 7. Governing Law and Jurisdiction. This Agreement will be governed and construed under the laws of the State of California without regard to choice of law principles, and the UN Convention on Contracts for the International Sale of Goods does not apply to this Agreement. The courts of California shall have exclusive jurisdiction of any dispute arising out of this Agreement. ### Meta Llama 3 Acceptable Use Policy Meta is committed to promoting safe and fair use of its tools and features, including Meta Llama 3. If you access or use Meta Llama 3, you agree to this Acceptable Use Policy (“Policy”). The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy](https://llama.meta.com/llama3/use-policy) #### Prohibited Uses We want everyone to use Meta Llama 3 safely and responsibly. You agree you will not use, or allow others to use, Meta Llama 3 to: 1. Violate the law or others’ rights, including to: 1. Engage in, promote, generate, contribute to, encourage, plan, incite, or further illegal or unlawful activity or content, such as: 1. Violence or terrorism 2. Exploitation or harm to children, including the solicitation, creation, acquisition, or dissemination of child exploitative content or failure to report Child Sexual Abuse Material 3. Human trafficking, exploitation, and sexual violence 4. The illegal distribution of information or materials to minors, including obscene materials, or failure to employ legally required age-gating in connection with such information or materials. 5. Sexual solicitation 6. Any other criminal activity 2. Engage in, promote, incite, or facilitate the harassment, abuse, threatening, or bullying of individuals or groups of individuals 3. Engage in, promote, incite, or facilitate discrimination or other unlawful or harmful conduct in the provision of employment, employment benefits, credit, housing, other economic benefits, or other essential goods and services 4. Engage in the unauthorized or unlicensed practice of any profession including, but not limited to, financial, legal, medical/health, or related professional practices 5. Collect, process, disclose, generate, or infer health, demographic, or other sensitive personal or private information about individuals without rights and consents required by applicable laws 6. Engage in or facilitate any action or generate any content that infringes, misappropriates, or otherwise violates any third-party rights, including the outputs or results of any products or services using the Llama Materials 7. Create, generate, or facilitate the creation of malicious code, malware, computer viruses or do anything else that could disable, overburden, interfere with or impair the proper working, integrity, operation or appearance of a website or computer system 2. Engage in, promote, incite, facilitate, or assist in the planning or development of activities that present a risk of death or bodily harm to individuals, including use of Meta Llama 3 related to the following: 1. Military, warfare, nuclear industries or applications, espionage, use for materials or activities that are subject to the International Traffic Arms Regulations (ITAR) maintained by the United States Department of State 2. Guns and illegal weapons (including weapon development) 3. Illegal drugs and regulated/controlled substances 4. Operation of critical infrastructure, transportation technologies, or heavy machinery 5. Self-harm or harm to others, including suicide, cutting, and eating disorders 6. Any content intended to incite or promote violence, abuse, or any infliction of bodily harm to an individual 3. Intentionally deceive or mislead others, including use of Meta Llama 3 related to the following: 1. Generating, promoting, or furthering fraud or the creation or promotion of disinformation 2. Generating, promoting, or furthering defamatory content, including the creation of defamatory statements, images, or other content 3. Generating, promoting, or further distributing spam 4. Impersonating another individual without consent, authorization, or legal right 5. Representing that the use of Meta Llama 3 or outputs are human-generated 6. Generating or facilitating false online engagement, including fake reviews and other means of fake online engagement 4. Fail to appropriately disclose to end users any known dangers of your AI system Please report any violation of this Policy, software “bug,” or other problems that could lead to a violation of this Policy through one of the following means: * Reporting issues with the model: [https://github.com/meta-llama/llama3](https://github.com/meta-llama/llama3) * Reporting risky content generated by the model: developers.facebook.com/llama_output_feedback * Reporting bugs and security concerns: facebook.com/whitehat/info * Reporting violations of the Acceptable Use Policy or unlicensed uses of Meta Llama 3: LlamaUseReport@meta.com extra_gated_fields: First Name: text Last Name: text Date of birth: date_picker Country: country Affiliation: text geo: ip_location By clicking Submit below I accept the terms of the license and acknowledge that the information I provide will be collected stored processed and shared in accordance with the Meta Privacy Policy: checkbox extra_gated_description: The information you provide will be collected, stored, processed and shared in accordance with the [Meta Privacy Policy](https://www.facebook.com/privacy/policy/). extra_gated_button_content: Submit --- ## Model Details Meta developed and released the Meta Llama 3 family of large language models (LLMs), a collection of pretrained and instruction tuned generative text models in 8 and 70B sizes. The Llama 3 instruction tuned models are optimized for dialogue use cases and outperform many of the available open source chat models on common industry benchmarks. Further, in developing these models, we took great care to optimize helpfulness and safety. **Model developers** Meta **Variations** Llama 3 comes in two sizes — 8B and 70B parameters — in pre-trained and instruction tuned variants. **Input** Models input text only. **Output** Models generate text and code only. **Model Architecture** Llama 3 is an auto-regressive language model that uses an optimized transformer architecture. The tuned versions use supervised fine-tuning (SFT) and reinforcement learning with human feedback (RLHF) to align with human preferences for helpfulness and safety. <table> <tr> <td> </td> <td><strong>Training Data</strong> </td> <td><strong>Params</strong> </td> <td><strong>Context length</strong> </td> <td><strong>GQA</strong> </td> <td><strong>Token count</strong> </td> <td><strong>Knowledge cutoff</strong> </td> </tr> <tr> <td rowspan="2" >Llama 3 </td> <td rowspan="2" >A new mix of publicly available online data. </td> <td>8B </td> <td>8k </td> <td>Yes </td> <td rowspan="2" >15T+ </td> <td>March, 2023 </td> </tr> <tr> <td>70B </td> <td>8k </td> <td>Yes </td> <td>December, 2023 </td> </tr> </table> **Llama 3 family of models**. Token counts refer to pretraining data only. Both the 8 and 70B versions use Grouped-Query Attention (GQA) for improved inference scalability. **Model Release Date** April 18, 2024. **Status** This is a static model trained on an offline dataset. Future versions of the tuned models will be released as we improve model safety with community feedback. **License** A custom commercial license is available at: [https://llama.meta.com/llama3/license](https://llama.meta.com/llama3/license) Where to send questions or comments about the model Instructions on how to provide feedback or comments on the model can be found in the model [README](https://github.com/meta-llama/llama3). For more technical information about generation parameters and recipes for how to use Llama 3 in applications, please go [here](https://github.com/meta-llama/llama-recipes). ## Intended Use **Intended Use Cases** Llama 3 is intended for commercial and research use in English. Instruction tuned models are intended for assistant-like chat, whereas pretrained models can be adapted for a variety of natural language generation tasks. **Out-of-scope** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in any other way that is prohibited by the Acceptable Use Policy and Llama 3 Community License. Use in languages other than English**. **Note: Developers may fine-tune Llama 3 models for languages beyond English provided they comply with the Llama 3 Community License and the Acceptable Use Policy. ## How to use This repository contains two versions of Meta-Llama-3-8B-Instruct, for use with transformers and with the original `llama3` codebase. ### Use with transformers See the snippet below for usage with Transformers: ```python >>> import transformers >>> import torch >>> model_id = "meta-llama/Meta-Llama-3-70B" >>> pipeline = transformers.pipeline( "text-generation", model=model_id, model_kwargs={"torch_dtype": torch.bfloat16}, device_map="auto" ) >>> pipeline("Hey how are you doing today?") ``` ### Use with `llama3` Please, follow the instructions in the [repository](https://github.com/meta-llama/llama3). To download Original checkpoints, see the example command below leveraging `huggingface-cli`: ``` huggingface-cli download meta-llama/Meta-Llama-3-70B --include "original/*" --local-dir Meta-Llama-3-70B ``` For Hugging Face support, we recommend using transformers or TGI, but a similar command works. ## Hardware and Software **Training Factors** We used custom training libraries, Meta's Research SuperCluster, and production clusters for pretraining. Fine-tuning, annotation, and evaluation were also performed on third-party cloud compute. **Carbon Footprint Pretraining utilized a cumulative** 7.7M GPU hours of computation on hardware of type H100-80GB (TDP of 700W). Estimated total emissions were 2290 tCO2eq, 100% of which were offset by Meta’s sustainability program. <table> <tr> <td> </td> <td><strong>Time (GPU hours)</strong> </td> <td><strong>Power Consumption (W)</strong> </td> <td><strong>Carbon Emitted(tCO2eq)</strong> </td> </tr> <tr> <td>Llama 3 8B </td> <td>1.3M </td> <td>700 </td> <td>390 </td> </tr> <tr> <td>Llama 3 70B </td> <td>6.4M </td> <td>700 </td> <td>1900 </td> </tr> <tr> <td>Total </td> <td>7.7M </td> <td> </td> <td>2290 </td> </tr> </table> **CO2 emissions during pre-training**. Time: total GPU time required for training each model. Power Consumption: peak power capacity per GPU device for the GPUs used adjusted for power usage efficiency. 100% of the emissions are directly offset by Meta's sustainability program, and because we are openly releasing these models, the pretraining costs do not need to be incurred by others. ## Training Data **Overview** Llama 3 was pretrained on over 15 trillion tokens of data from publicly available sources. The fine-tuning data includes publicly available instruction datasets, as well as over 10M human-annotated examples. Neither the pretraining nor the fine-tuning datasets include Meta user data. **Data Freshness** The pretraining data has a cutoff of March 2023 for the 7B and December 2023 for the 70B models respectively. ## Benchmarks In this section, we report the results for Llama 3 models on standard automatic benchmarks. For all the evaluations, we use our internal evaluations library. For details on the methodology see [here](https://github.com/meta-llama/llama3/blob/main/eval_methodology.md). ### Base pretrained models <table> <tr> <td><strong>Category</strong> </td> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama2 7B</strong> </td> <td><strong>Llama2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama2 70B</strong> </td> </tr> <tr> <td rowspan="6" >General </td> <td>MMLU (5-shot) </td> <td>66.6 </td> <td>45.7 </td> <td>53.8 </td> <td>79.5 </td> <td>69.7 </td> </tr> <tr> <td>AGIEval English (3-5 shot) </td> <td>45.9 </td> <td>28.8 </td> <td>38.7 </td> <td>63.0 </td> <td>54.8 </td> </tr> <tr> <td>CommonSenseQA (7-shot) </td> <td>72.6 </td> <td>57.6 </td> <td>67.6 </td> <td>83.8 </td> <td>78.7 </td> </tr> <tr> <td>Winogrande (5-shot) </td> <td>76.1 </td> <td>73.3 </td> <td>75.4 </td> <td>83.1 </td> <td>81.8 </td> </tr> <tr> <td>BIG-Bench Hard (3-shot, CoT) </td> <td>61.1 </td> <td>38.1 </td> <td>47.0 </td> <td>81.3 </td> <td>65.7 </td> </tr> <tr> <td>ARC-Challenge (25-shot) </td> <td>78.6 </td> <td>53.7 </td> <td>67.6 </td> <td>93.0 </td> <td>85.3 </td> </tr> <tr> <td>Knowledge reasoning </td> <td>TriviaQA-Wiki (5-shot) </td> <td>78.5 </td> <td>72.1 </td> <td>79.6 </td> <td>89.7 </td> <td>87.5 </td> </tr> <tr> <td rowspan="4" >Reading comprehension </td> <td>SQuAD (1-shot) </td> <td>76.4 </td> <td>72.2 </td> <td>72.1 </td> <td>85.6 </td> <td>82.6 </td> </tr> <tr> <td>QuAC (1-shot, F1) </td> <td>44.4 </td> <td>39.6 </td> <td>44.9 </td> <td>51.1 </td> <td>49.4 </td> </tr> <tr> <td>BoolQ (0-shot) </td> <td>75.7 </td> <td>65.5 </td> <td>66.9 </td> <td>79.0 </td> <td>73.1 </td> </tr> <tr> <td>DROP (3-shot, F1) </td> <td>58.4 </td> <td>37.9 </td> <td>49.8 </td> <td>79.7 </td> <td>70.2 </td> </tr> </table> ### Instruction tuned models <table> <tr> <td><strong>Benchmark</strong> </td> <td><strong>Llama 3 8B</strong> </td> <td><strong>Llama 2 7B</strong> </td> <td><strong>Llama 2 13B</strong> </td> <td><strong>Llama 3 70B</strong> </td> <td><strong>Llama 2 70B</strong> </td> </tr> <tr> <td>MMLU (5-shot) </td> <td>68.4 </td> <td>34.1 </td> <td>47.8 </td> <td>82.0 </td> <td>52.9 </td> </tr> <tr> <td>GPQA (0-shot) </td> <td>34.2 </td> <td>21.7 </td> <td>22.3 </td> <td>39.5 </td> <td>21.0 </td> </tr> <tr> <td>HumanEval (0-shot) </td> <td>62.2 </td> <td>7.9 </td> <td>14.0 </td> <td>81.7 </td> <td>25.6 </td> </tr> <tr> <td>GSM-8K (8-shot, CoT) </td> <td>79.6 </td> <td>25.7 </td> <td>77.4 </td> <td>93.0 </td> <td>57.5 </td> </tr> <tr> <td>MATH (4-shot, CoT) </td> <td>30.0 </td> <td>3.8 </td> <td>6.7 </td> <td>50.4 </td> <td>11.6 </td> </tr> </table> ### Responsibility & Safety We believe that an open approach to AI leads to better, safer products, faster innovation, and a bigger overall market. We are committed to Responsible AI development and took a series of steps to limit misuse and harm and support the open source community. Foundation models are widely capable technologies that are built to be used for a diverse range of applications. They are not designed to meet every developer preference on safety levels for all use cases, out-of-the-box, as those by their nature will differ across different applications. Rather, responsible LLM-application deployment is achieved by implementing a series of safety best practices throughout the development of such applications, from the model pre-training, fine-tuning and the deployment of systems composed of safeguards to tailor the safety needs specifically to the use case and audience. As part of the Llama 3 release, we updated our [Responsible Use Guide](https://llama.meta.com/responsible-use-guide/) to outline the steps and best practices for developers to implement model and system level safety for their application. We also provide a set of resources including [Meta Llama Guard 2](https://llama.meta.com/purple-llama/) and [Code Shield](https://llama.meta.com/purple-llama/) safeguards. These tools have proven to drastically reduce residual risks of LLM Systems, while maintaining a high level of helpfulness. We encourage developers to tune and deploy these safeguards according to their needs and we provide a [reference implementation](https://github.com/meta-llama/llama-recipes/tree/main/recipes/responsible_ai) to get you started. #### Llama 3-Instruct As outlined in the Responsible Use Guide, some trade-off between model helpfulness and model alignment is likely unavoidable. Developers should exercise discretion about how to weigh the benefits of alignment and helpfulness for their specific use case and audience. Developers should be mindful of residual risks when using Llama models and leverage additional safety tools as needed to reach the right safety bar for their use case. <span style="text-decoration:underline;">Safety</span> For our instruction tuned model, we conducted extensive red teaming exercises, performed adversarial evaluations and implemented safety mitigations techniques to lower residual risks. As with any Large Language Model, residual risks will likely remain and we recommend that developers assess these risks in the context of their use case. In parallel, we are working with the community to make AI safety benchmark standards transparent, rigorous and interpretable. <span style="text-decoration:underline;">Refusals</span> In addition to residual risks, we put a great emphasis on model refusals to benign prompts. Over-refusing not only can impact the user experience but could even be harmful in certain contexts as well. We’ve heard the feedback from the developer community and improved our fine tuning to ensure that Llama 3 is significantly less likely to falsely refuse to answer prompts than Llama 2. We built internal benchmarks and developed mitigations to limit false refusals making Llama 3 our most helpful model to date. #### Responsible release In addition to responsible use considerations outlined above, we followed a rigorous process that requires us to take extra measures against misuse and critical risks before we make our release decision. Misuse If you access or use Llama 3, you agree to the Acceptable Use Policy. The most recent copy of this policy can be found at [https://llama.meta.com/llama3/use-policy/](https://llama.meta.com/llama3/use-policy/). #### Critical risks <span style="text-decoration:underline;">CBRNE</span> (Chemical, Biological, Radiological, Nuclear, and high yield Explosives) We have conducted a two fold assessment of the safety of the model in this area: * Iterative testing during model training to assess the safety of responses related to CBRNE threats and other adversarial risks. * Involving external CBRNE experts to conduct an uplift test assessing the ability of the model to accurately provide expert knowledge and reduce barriers to potential CBRNE misuse, by reference to what can be achieved using web search (without the model). ### <span style="text-decoration:underline;">Cyber Security </span> We have evaluated Llama 3 with CyberSecEval, Meta’s cybersecurity safety eval suite, measuring Llama 3’s propensity to suggest insecure code when used as a coding assistant, and Llama 3’s propensity to comply with requests to help carry out cyber attacks, where attacks are defined by the industry standard MITRE ATT&CK cyber attack ontology. On our insecure coding and cyber attacker helpfulness tests, Llama 3 behaved in the same range or safer than models of [equivalent coding capability](https://huggingface.co/spaces/facebook/CyberSecEval). ### <span style="text-decoration:underline;">Child Safety</span> Child Safety risk assessments were conducted using a team of experts, to assess the model’s capability to produce outputs that could result in Child Safety risks and inform on any necessary and appropriate risk mitigations via fine tuning. We leveraged those expert red teaming sessions to expand the coverage of our evaluation benchmarks through Llama 3 model development. For Llama 3, we conducted new in-depth sessions using objective based methodologies to assess the model risks along multiple attack vectors. We also partnered with content specialists to perform red teaming exercises assessing potentially violating content while taking account of market specific nuances or experiences. ### Community Generative AI safety requires expertise and tooling, and we believe in the strength of the open community to accelerate its progress. We are active members of open consortiums, including the AI Alliance, Partnership in AI and MLCommons, actively contributing to safety standardization and transparency. We encourage the community to adopt taxonomies like the MLCommons Proof of Concept evaluation to facilitate collaboration and transparency on safety and content evaluations. Our Purple Llama tools are open sourced for the community to use and widely distributed across ecosystem partners including cloud service providers. We encourage community contributions to our [Github repository](https://github.com/meta-llama/PurpleLlama). Finally, we put in place a set of resources including an [output reporting mechanism](https://developers.facebook.com/llama_output_feedback) and [bug bounty program](https://www.facebook.com/whitehat) to continuously improve the Llama technology with the help of the community. ## Ethical Considerations and Limitations The core values of Llama 3 are openness, inclusivity and helpfulness. It is meant to serve everyone, and to work for a wide range of use cases. It is thus designed to be accessible to people across many different backgrounds, experiences and perspectives. Llama 3 addresses users and their needs as they are, without insertion unnecessary judgment or normativity, while reflecting the understanding that even content that may appear problematic in some cases can serve valuable purposes in others. It respects the dignity and autonomy of all users, especially in terms of the values of free thought and expression that power innovation and progress. But Llama 3 is a new technology, and like any new technology, there are risks associated with its use. Testing conducted to date has been in English, and has not covered, nor could it cover, all scenarios. For these reasons, as with all LLMs, Llama 3’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate, biased or other objectionable responses to user prompts. Therefore, before deploying any applications of Llama 3 models, developers should perform safety testing and tuning tailored to their specific applications of the model. As outlined in the Responsible Use Guide, we recommend incorporating [Purple Llama](https://github.com/facebookresearch/PurpleLlama) solutions into your workflows and specifically [Llama Guard](https://ai.meta.com/research/publications/llama-guard-llm-based-input-output-safeguard-for-human-ai-conversations/) which provides a base model to filter input and output prompts to layer system-level safety on top of model-level safety. Please see the Responsible Use Guide available at [http://llama.meta.com/responsible-use-guide](http://llama.meta.com/responsible-use-guide) ## Citation instructions @article{llama3modelcard, title={Llama 3 Model Card}, author={AI@Meta}, year={2024}, url = {https://github.com/meta-llama/llama3/blob/main/MODEL_CARD.md} } ## Contributors Aaditya Singh; Aaron Grattafiori; Abhimanyu Dubey; Abhinav Jauhri; Abhinav Pandey; Abhishek Kadian; Adam Kelsey; Adi Gangidi; Ahmad Al-Dahle; Ahuva Goldstand; Aiesha Letman; Ajay Menon; Akhil Mathur; Alan Schelten; Alex Vaughan; Amy Yang; Andrei Lupu; Andres Alvarado; Andrew Gallagher; Andrew Gu; Andrew Ho; Andrew Poulton; Andrew Ryan; Angela Fan; Ankit Ramchandani; Anthony Hartshorn; Archi Mitra; Archie Sravankumar; Artem Korenev; Arun Rao; Ashley Gabriel; Ashwin Bharambe; Assaf Eisenman; Aston Zhang; Aurelien Rodriguez; Austen Gregerson; Ava Spataru; Baptiste Roziere; Ben Maurer; Benjamin Leonhardi; Bernie Huang; Bhargavi Paranjape; Bing Liu; Binh Tang; Bobbie Chern; Brani Stojkovic; Brian Fuller; Catalina Mejia Arenas; Chao Zhou; Charlotte Caucheteux; Chaya Nayak; Ching-Hsiang Chu; Chloe Bi; Chris Cai; Chris Cox; Chris Marra; Chris McConnell; Christian Keller; Christoph Feichtenhofer; Christophe Touret; Chunyang Wu; Corinne Wong; Cristian Canton Ferrer; Damien Allonsius; Daniel Kreymer; Daniel Haziza; Daniel Li; Danielle Pintz; Danny Livshits; Danny Wyatt; David Adkins; David Esiobu; David Xu; Davide Testuggine; Delia David; Devi Parikh; Dhruv Choudhary; Dhruv Mahajan; Diana Liskovich; Diego Garcia-Olano; Diego Perino; Dieuwke Hupkes; Dingkang Wang; Dustin Holland; Egor Lakomkin; Elina Lobanova; Xiaoqing Ellen Tan; Emily Dinan; Eric Smith; Erik Brinkman; Esteban Arcaute; Filip Radenovic; Firat Ozgenel; Francesco Caggioni; Frank Seide; Frank Zhang; Gabriel Synnaeve; Gabriella Schwarz; Gabrielle Lee; Gada Badeer; Georgia Anderson; Graeme Nail; Gregoire Mialon; Guan Pang; Guillem Cucurell; Hailey Nguyen; Hannah Korevaar; Hannah Wang; Haroun Habeeb; Harrison Rudolph; Henry Aspegren; Hu Xu; Hugo Touvron; Iga Kozlowska; Igor Molybog; Igor Tufanov; Iliyan Zarov; Imanol Arrieta Ibarra; Irina-Elena Veliche; Isabel Kloumann; Ishan Misra; Ivan Evtimov; Jacob Xu; Jade Copet; Jake Weissman; Jan Geffert; Jana Vranes; Japhet Asher; Jason Park; Jay Mahadeokar; Jean-Baptiste Gaya; Jeet Shah; Jelmer van der Linde; Jennifer Chan; Jenny Hong; Jenya Lee; Jeremy Fu; Jeremy Teboul; Jianfeng Chi; Jianyu Huang; Jie Wang; Jiecao Yu; Joanna Bitton; Joe Spisak; Joelle Pineau; Jon Carvill; Jongsoo Park; Joseph Rocca; Joshua Johnstun; Junteng Jia; Kalyan Vasuden Alwala; Kam Hou U; Kate Plawiak; Kartikeya Upasani; Kaushik Veeraraghavan; Ke Li; Kenneth Heafield; Kevin Stone; Khalid El-Arini; Krithika Iyer; Kshitiz Malik; Kuenley Chiu; Kunal Bhalla; Kyle Huang; Lakshya Garg; Lauren Rantala-Yeary; Laurens van der Maaten; Lawrence Chen; Leandro Silva; Lee Bell; Lei Zhang; Liang Tan; Louis Martin; Lovish Madaan; Luca Wehrstedt; Lukas Blecher; Luke de Oliveira; Madeline Muzzi; Madian Khabsa; Manav Avlani; Mannat Singh; Manohar Paluri; Mark Zuckerberg; Marcin Kardas; Martynas Mankus; Mathew Oldham; Mathieu Rita; Matthew Lennie; Maya Pavlova; Meghan Keneally; Melanie Kambadur; Mihir Patel; Mikayel Samvelyan; Mike Clark; Mike Lewis; Min Si; Mitesh Kumar Singh; Mo Metanat; Mona Hassan; Naman Goyal; Narjes Torabi; Nicolas Usunier; Nikolay Bashlykov; Nikolay Bogoychev; Niladri Chatterji; Ning Dong; Oliver Aobo Yang; Olivier Duchenne; Onur Celebi; Parth Parekh; Patrick Alrassy; Paul Saab; Pavan Balaji; Pedro Rittner; Pengchuan Zhang; Pengwei Li; Petar Vasic; Peter Weng; Polina Zvyagina; Prajjwal Bhargava; Pratik Dubal; Praveen Krishnan; Punit Singh Koura; Qing He; Rachel Rodriguez; Ragavan Srinivasan; Rahul Mitra; Ramon Calderer; Raymond Li; Robert Stojnic; Roberta Raileanu; Robin Battey; Rocky Wang; Rohit Girdhar; Rohit Patel; Romain Sauvestre; Ronnie Polidoro; Roshan Sumbaly; Ross Taylor; Ruan Silva; Rui Hou; Rui Wang; Russ Howes; Ruty Rinott; Saghar Hosseini; Sai Jayesh Bondu; Samyak Datta; Sanjay Singh; Sara Chugh; Sargun Dhillon; Satadru Pan; Sean Bell; Sergey Edunov; Shaoliang Nie; Sharan Narang; Sharath Raparthy; Shaun Lindsay; Sheng Feng; Sheng Shen; Shenghao Lin; Shiva Shankar; Shruti Bhosale; Shun Zhang; Simon Vandenhende; Sinong Wang; Seohyun Sonia Kim; Soumya Batra; Sten Sootla; Steve Kehoe; Suchin Gururangan; Sumit Gupta; Sunny Virk; Sydney Borodinsky; Tamar Glaser; Tamar Herman; Tamara Best; Tara Fowler; Thomas Georgiou; Thomas Scialom; Tianhe Li; Todor Mihaylov; Tong Xiao; Ujjwal Karn; Vedanuj Goswami; Vibhor Gupta; Vignesh Ramanathan; Viktor Kerkez; Vinay Satish Kumar; Vincent Gonguet; Vish Vogeti; Vlad Poenaru; Vlad Tiberiu Mihailescu; Vladan Petrovic; Vladimir Ivanov; Wei Li; Weiwei Chu; Wenhan Xiong; Wenyin Fu; Wes Bouaziz; Whitney Meers; Will Constable; Xavier Martinet; Xiaojian Wu; Xinbo Gao; Xinfeng Xie; Xuchao Jia; Yaelle Goldschlag; Yann LeCun; Yashesh Gaur; Yasmine Babaei; Ye Qi; Yenda Li; Yi Wen; Yiwen Song; Youngjin Nam; Yuchen Hao; Yuchen Zhang; Yun Wang; Yuning Mao; Yuzi He; Zacharie Delpierre Coudert; Zachary DeVito; Zahra Hankir; Zhaoduo Wen; Zheng Yan; Zhengxing Chen; Zhenyu Yang; Zoe Papakipos

bhadresh-savani/distilbert-base-uncased-emotion

bhadresh-savani

transformers

text-classification

--- language: - en license: apache-2.0 tags: - text-classification - emotion - pytorch datasets: - emotion metrics: - Accuracy, F1 Score thumbnail: https://avatars3.githubusercontent.com/u/32437151?s=460&u=4ec59abc8d21d5feea3dab323d23a5860e6996a4&v=4 model-index: - name: bhadresh-savani/distilbert-base-uncased-emotion results: - task: type: text-classification name: Text Classification dataset: name: emotion type: emotion config: default split: test metrics: - type: accuracy value: 0.927 name: Accuracy verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYzQxOGRmMjFlZThmZWViNjNmNGMzMTdjMGNjYjg1YWUzOTI0ZDlmYjRhYWMzMDA3Yjg2N2FiMTdmMzk0ZjJkOSIsInZlcnNpb24iOjF9.mOqr-hgNrnle7WCPy3Mo7M3fITFppn5gjpNagGMf_TZfB6VZnPKfZ51UkNFQlBtUlcm0U8vwPkF79snxwvCoDw - type: precision value: 0.8880230732280744 name: Precision Macro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYjZiN2NjNTkyN2M3ZWM2ZDZiNDk1OWZhN2FmNTAwZDIzMmQ3NTU2Yjk2MTgyNjJmMTNjYTYzOTc1NDdhYTljYSIsInZlcnNpb24iOjF9.0rWHmCZ2PyZ5zYkSeb_tFdQG9CHS5PdpOZ9kOfrIzEXyZ968daayaOJi2d6iO84fnauE5hZiIAUPsx24Vr4nBA - type: precision value: 0.927 name: Precision Micro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiZmRhNWM1NDQ4ZjkyYjAxYjQ5MzQzMDA1ZDIzYWU3YTE4NTI2ZTMwYWI2ZWQ4NzQ3YzJkODYzMmZhZDI1NGRlNCIsInZlcnNpb24iOjF9.NlII1s42Mr_DMzPEoR0ntyh5cDW0405TxVkWhCgXLJTFAdnivH54-zZY4av1U5jHPTeXeWwZrrrbMwHCRBkoCw - type: precision value: 0.9272902840835793 name: Precision Weighted verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiODhkNmM5NmYyMzA4MjkwOTllZDgyMDQ1NzZkN2QzOTAyOTMyNGFlZTU4NzM5NmM5NWQ1YmUxYmRmNjA5YjhhNCIsInZlcnNpb24iOjF9.oIn1KT-BOpFNLXiKL29frMvgHhWZMHWc9Q5WgeR7UaMEO7smkK8J3j5HAMy17Ktjv2dh783-f76N6gyJ_NewCg - type: recall value: 0.8790126653780703 name: Recall Macro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiYjhlNzczNDY2NDVlM2UwMjAzOWQxYTAyNWZkNGZlYmNjODNiZTEzMTcxNTE3MTAxNjNkOTFiMmRiMzViMzJmZiIsInZlcnNpb24iOjF9.AXp7omMuUZFJ6mzAVTQPMke7QoUtoi4RJSSE7Xbnp2pNi7y-JtznKdm---l6RfqcHPlI0jWr7TVGoFsWZ64YAg - type: recall value: 0.927 name: Recall Micro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMjEyYmZiZDQ4MzM1ZmQ2ZmJhZWU4OTVkNmViYjA5NzhiN2MxODE0MzUxZTliZTk0MzViZDAyNGU4MDFjYjM1MSIsInZlcnNpb24iOjF9.9lazxLXbPOdwhqoYtIudwRwjfNVZnUu7KvGRklRP_RAoQStAzgmWMIrT3ckX_d5_6bKZH9fIdujUn5Qz-baKBw - type: recall value: 0.927 name: Recall Weighted verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMWVhMzY0YTA4YmQzYTg4YTBiMzQ5YzRiZWJhMjM1NjUzZGQxZmQ5M2NkZDcyNTQ0ZmJjN2NkY2ZiYjg0OWI0ZCIsInZlcnNpb24iOjF9.QgTv726WCTyvrEct0NM8Zpc3vUnDbIwCor9EH941-zpJtuWr-xpdZzYZFJfILkVA0UUn1y6Jz_ABfkfBeyZTBg - type: f1 value: 0.8825061528287809 name: F1 Macro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNzQzZTJkMDAwOTUwMzY3ZjI2MjIxYjlmZTg3YTdhNTc4ZjYyMmQ2NDQzM2FmYzk3OGEzNjhhMTk3NTQ3OTlhNyIsInZlcnNpb24iOjF9.hSln1KfKm0plK7Qao9vlubFtAl1M7_UYHNM6La9gEZlW_apnU1Mybz03GT2XZORgOVPe9JmgygvZByxQhpsYBw - type: f1 value: 0.927 name: F1 Micro verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiNzljODQ3NjE3MDRkODE3ZjFlZmY5MjYyOGJlNDQ4YzdlZGRiMTI5OGZiZWM2ODkyZjMyZWQ3MTkzYWU5YThkOCIsInZlcnNpb24iOjF9.7qfBw39fv22jSIJoY71DkOVr9eBB-srhqSi09bCcUC7Huok4O2Z_vB7gO_Rahh9sFgKVu1ZATusjTmOLQr0fBw - type: f1 value: 0.926876082854655 name: F1 Weighted verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMjJhN2UzODgxOWQ0Y2E3YTcwZTQxMDE0ZWRmYThjOWVhYWQ1YjBhMzk0YWUxNzE2ZjFhNWM5ZmE2ZmI1YTczYSIsInZlcnNpb24iOjF9.nZW0dBdLmh_FgNw6GaITvSJFX-2C_Iku3NanU8Rip7FSiRHozKPAjothdQh9MWQnq158ZZGPPVIjtyIvuTSqCw - type: loss value: 0.17403268814086914 name: loss verified: true verifyToken: eyJhbGciOiJFZERTQSIsInR5cCI6IkpXVCJ9.eyJoYXNoIjoiMTVjZmFiOGQwZGY1OTU5YWFkNGZjMTlhOGI4NjE3MGI4ZDhkODcxYmJiYTQ3NWNmMWM0ODUyZDI1MThkYTY3ZSIsInZlcnNpb24iOjF9.OYz5BI3Lz8LgjAqVnD6NcrG3UAG0D3wjKJ7G5298RRGaNpb621ycisG_7UYiWixY7e2RJafkfRiplmkdczIFDQ --- # Distilbert-base-uncased-emotion ## Model description: [Distilbert](https://arxiv.org/abs/1910.01108) is created with knowledge distillation during the pre-training phase which reduces the size of a BERT model by 40%, while retaining 97% of its language understanding. It's smaller, faster than Bert and any other Bert-based model. [Distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) finetuned on the emotion dataset using HuggingFace Trainer with below Hyperparameters ``` learning rate 2e-5, batch size 64, num_train_epochs=8, ``` ## Model Performance Comparision on Emotion Dataset from Twitter: | Model | Accuracy | F1 Score | Test Sample per Second | | --- | --- | --- | --- | | [Distilbert-base-uncased-emotion](https://huggingface.co/bhadresh-savani/distilbert-base-uncased-emotion) | 93.8 | 93.79 | 398.69 | | [Bert-base-uncased-emotion](https://huggingface.co/bhadresh-savani/bert-base-uncased-emotion) | 94.05 | 94.06 | 190.152 | | [Roberta-base-emotion](https://huggingface.co/bhadresh-savani/roberta-base-emotion) | 93.95 | 93.97| 195.639 | | [Albert-base-v2-emotion](https://huggingface.co/bhadresh-savani/albert-base-v2-emotion) | 93.6 | 93.65 | 182.794 | ## How to Use the model: ```python from transformers import pipeline classifier = pipeline("text-classification",model='bhadresh-savani/distilbert-base-uncased-emotion', return_all_scores=True) prediction = classifier("I love using transformers. The best part is wide range of support and its easy to use", ) print(prediction) """ Output: [[ {'label': 'sadness', 'score': 0.0006792712374590337}, {'label': 'joy', 'score': 0.9959300756454468}, {'label': 'love', 'score': 0.0009452480007894337}, {'label': 'anger', 'score': 0.0018055217806249857}, {'label': 'fear', 'score': 0.00041110432357527316}, {'label': 'surprise', 'score': 0.0002288572577526793} ]] """ ``` ## Dataset: [Twitter-Sentiment-Analysis](https://huggingface.co/nlp/viewer/?dataset=emotion). ## Training procedure [Colab Notebook](https://github.com/bhadreshpsavani/ExploringSentimentalAnalysis/blob/main/SentimentalAnalysisWithDistilbert.ipynb) ## Eval results ```json { 'test_accuracy': 0.938, 'test_f1': 0.937932884041714, 'test_loss': 0.1472451239824295, 'test_mem_cpu_alloc_delta': 0, 'test_mem_cpu_peaked_delta': 0, 'test_mem_gpu_alloc_delta': 0, 'test_mem_gpu_peaked_delta': 163454464, 'test_runtime': 5.0164, 'test_samples_per_second': 398.69 } ``` ## Reference: * [Natural Language Processing with Transformer By Lewis Tunstall, Leandro von Werra, Thomas Wolf](https://learning.oreilly.com/library/view/natural-language-processing/9781098103231/)

timm/deit_base_distilled_patch16_224.fb_in1k

timm

timm

image-classification

--- license: apache-2.0 library_name: timm tags: - image-classification - timm datasets: - imagenet-1k --- # Model card for deit_base_distilled_patch16_224.fb_in1k A DeiT image classification model. Trained on ImageNet-1k using distillation tokens by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 87.3 - GMACs: 17.7 - Activations (M): 24.0 - Image size: 224 x 224 - **Papers:** - Training data-efficient image transformers & distillation through attention: https://arxiv.org/abs/2012.12877 - **Original:** https://github.com/facebookresearch/deit - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('deit_base_distilled_patch16_224.fb_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'deit_base_distilled_patch16_224.fb_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 198, 768) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @InProceedings{pmlr-v139-touvron21a, title = {Training data-efficient image transformers & distillation through attention}, author = {Touvron, Hugo and Cord, Matthieu and Douze, Matthijs and Massa, Francisco and Sablayrolles, Alexandre and Jegou, Herve}, booktitle = {International Conference on Machine Learning}, pages = {10347--10357}, year = {2021}, volume = {139}, month = {July} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

CohereForAI/aya-23-35B

CohereForAI

transformers

text-generation

--- inference: false library_name: transformers language: - en - fr - de - es - it - pt - ja - ko - zh - ar - el - fa - pl - id - cs - he - hi - nl - ro - ru - tr - uk - vi license: cc-by-nc-4.0 --- # Model Card for Aya-23-35B **Try Aya 23** You can try out Aya 23 (35B) before downloading the weights in our hosted Hugging Face Space [here](https://huggingface.co/spaces/CohereForAI/aya-23). ## Model Summary Aya 23 is an open weights research release of an instruction fine-tuned model with highly advanced multilingual capabilities. Aya 23 focuses on pairing a highly performant pre-trained [Command family](https://huggingface.co/CohereForAI/c4ai-command-r-plus) of models with the recently released [Aya Collection](https://huggingface.co/datasets/CohereForAI/aya_collection). The result is a powerful multilingual large language model serving 23 languages. This model card corresponds to the 35-billion version of the Aya 23 model. We also released an 8-billion version which you can find [here](https://huggingface.co/CohereForAI/aya-23-8B). We cover 23 languages: Arabic, Chinese (simplified & traditional), Czech, Dutch, English, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Persian, Polish, Portuguese, Romanian, Russian, Spanish, Turkish, Ukrainian, and Vietnamese Developed by: [Cohere For AI](https://cohere.for.ai) and [Cohere](https://cohere.com/) - Point of Contact: Cohere For AI: [cohere.for.ai](https://cohere.for.ai/) - License: [CC-BY-NC](https://cohere.com/c4ai-cc-by-nc-license), requires also adhering to [C4AI's Acceptable Use Policy](https://docs.cohere.com/docs/c4ai-acceptable-use-policy) - Model: aya-23-35B - Model Size: 35 billion parameters ### Usage Please install transformers from the source repository that includes the necessary changes for this model ```python # pip install 'git+https://github.com/huggingface/transformers.git' from transformers import AutoTokenizer, AutoModelForCausalLM model_id = "CohereForAI/aya-23-35B" tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained(model_id) # Format message with the command-r-plus chat template messages = [{"role": "user", "content": "Anneme onu ne kadar sevdiğimi anlatan bir mektup yaz"}] input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt") ## <BOS_TOKEN><|START_OF_TURN_TOKEN|><|USER_TOKEN|>Anneme onu ne kadar sevdiğimi anlatan bir mektup yaz<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|> gen_tokens = model.generate( input_ids, max_new_tokens=100, do_sample=True, temperature=0.3, ) gen_text = tokenizer.decode(gen_tokens[0]) print(gen_text) ``` ### Example Notebook [This notebook](https://huggingface.co/CohereForAI/aya-23-35B/blob/main/Aya_23_notebook.ipynb) showcases a detailed use of Aya 23 (8B) including inference and fine-tuning with [QLoRA](https://huggingface.co/blog/4bit-transformers-bitsandbytes). ## Model Details **Input**: Models input text only. **Output**: Models generate text only. **Model Architecture**: Aya-23-35B is an auto-regressive language model that uses an optimized transformer architecture. After pretraining, this model is fine-tuned (IFT) to follow human instructions. **Languages covered**: The model is particularly optimized for multilinguality and supports the following languages: Arabic, Chinese (simplified & traditional), Czech, Dutch, English, French, German, Greek, Hebrew, Hindi, Indonesian, Italian, Japanese, Korean, Persian, Polish, Portuguese, Romanian, Russian, Spanish, Turkish, Ukrainian, and Vietnamese **Context length**: 8192 ### Evaluation <img src="benchmarks.png" alt="multilingual benchmarks" width="650" style="margin-left:'auto' margin-right:'auto' display:'block'"/> <img src="winrates.png" alt="average win rates" width="650" style="margin-left:'auto' margin-right:'auto' display:'block'"/> Please refer to the [Aya 23 technical report](https://cohere.com/research/papers/aya-command-23-8b-and-35b-technical-report-2024-05-23) for further details about the base model, data, instruction tuning, and evaluation. ### Model Card Contact For errors or additional questions about details in this model card, contact info@for.ai. ### Terms of Use We hope that the release of this model will make community-based research efforts more accessible, by releasing the weights of a highly performant multilingual model to researchers all over the world. This model is governed by a [CC-BY-NC](https://cohere.com/c4ai-cc-by-nc-license) License with an acceptable use addendum, and also requires adhering to [C4AI's Acceptable Use Policy](https://docs.cohere.com/docs/c4ai-acceptable-use-policy). ### Try the model today You can try Aya 23 in the Cohere [playground](https://dashboard.cohere.com/playground/chat) here. You can also use it in our dedicated Hugging Face Space [here](https://huggingface.co/spaces/CohereForAI/aya-23). ### Citation info ```bibtex @misc{aryabumi2024aya, title={Aya 23: Open Weight Releases to Further Multilingual Progress}, author={Viraat Aryabumi and John Dang and Dwarak Talupuru and Saurabh Dash and David Cairuz and Hangyu Lin and Bharat Venkitesh and Madeline Smith and Kelly Marchisio and Sebastian Ruder and Acyr Locatelli and Julia Kreutzer and Nick Frosst and Phil Blunsom and Marzieh Fadaee and Ahmet Üstün and Sara Hooker}, year={2024}, eprint={2405.15032}, archivePrefix={arXiv}, primaryClass={cs.CL} }

eugenesiow/msrn

eugenesiow

transformers

--- license: apache-2.0 tags: - super-image - image-super-resolution datasets: - eugenesiow/Div2k - eugenesiow/Set5 - eugenesiow/Set14 - eugenesiow/BSD100 - eugenesiow/Urban100 metrics: - pnsr - ssim --- # Multi-scale Residual Network for Image Super-Resolution (MSRN) MSRN model pre-trained on DIV2K (800 images training, augmented to 4000 images, 100 images validation) for 2x, 3x and 4x image super resolution. It was introduced in the paper [Multi-scale Residual Network for Image Super-Resolution](https://openaccess.thecvf.com/content_ECCV_2018/html/Juncheng_Li_Multi-scale_Residual_Network_ECCV_2018_paper.html) by Li et al. (2018) and first released in [this repository](https://github.com/MIVRC/MSRN-PyTorch). The goal of image super resolution is to restore a high resolution (HR) image from a single low resolution (LR) image. The image below shows the ground truth (HR), the bicubic upscaling x2 and model upscaling x2.  ## Model description The MSRN model proposes a feature extraction structure called the multi-scale residual block. This module can "adaptively detect image features at different scales" and "exploit the potential features of the image". ## Intended uses & limitations You can use the pre-trained models for upscaling your images 2x, 3x and 4x. You can also use the trainer to train a model on your own dataset. ### How to use The model can be used with the [super_image](https://github.com/eugenesiow/super-image) library: ```bash pip install super-image ``` Here is how to use a pre-trained model to upscale your image: ```python from super_image import MsrnModel, ImageLoader from PIL import Image import requests url = 'https://paperswithcode.com/media/datasets/Set5-0000002728-07a9793f_zA3bDjj.jpg' image = Image.open(requests.get(url, stream=True).raw) model = MsrnModel.from_pretrained('eugenesiow/msrn', scale=4) # scale 2, 3 and 4 models available inputs = ImageLoader.load_image(image) preds = model(inputs) ImageLoader.save_image(preds, './scaled_4x.png') # save the output 4x scaled image to `./scaled_4x.png` ImageLoader.save_compare(inputs, preds, './scaled_4x_compare.png') # save an output comparing the super-image with a bicubic scaling ``` [](https://colab.research.google.com/github/eugenesiow/super-image-notebooks/blob/master/notebooks/Upscale_Images_with_Pretrained_super_image_Models.ipynb "Open in Colab") ## Training data The models for 2x, 3x and 4x image super resolution were pretrained on [DIV2K](https://huggingface.co/datasets/eugenesiow/Div2k), a dataset of 800 high-quality (2K resolution) images for training, augmented to 4000 images and uses a dev set of 100 validation images (images numbered 801 to 900). ## Training procedure ### Preprocessing We follow the pre-processing and training method of [Wang et al.](https://arxiv.org/abs/2104.07566). Low Resolution (LR) images are created by using bicubic interpolation as the resizing method to reduce the size of the High Resolution (HR) images by x2, x3 and x4 times. During training, RGB patches with size of 64×64 from the LR input are used together with their corresponding HR patches. Data augmentation is applied to the training set in the pre-processing stage where five images are created from the four corners and center of the original image. We need the huggingface [datasets](https://huggingface.co/datasets?filter=task_ids:other-other-image-super-resolution) library to download the data: ```bash pip install datasets ``` The following code gets the data and preprocesses/augments the data. ```python from datasets import load_dataset from super_image.data import EvalDataset, TrainDataset, augment_five_crop augmented_dataset = load_dataset('eugenesiow/Div2k', 'bicubic_x4', split='train')\ .map(augment_five_crop, batched=True, desc="Augmenting Dataset") # download and augment the data with the five_crop method train_dataset = TrainDataset(augmented_dataset) # prepare the train dataset for loading PyTorch DataLoader eval_dataset = EvalDataset(load_dataset('eugenesiow/Div2k', 'bicubic_x4', split='validation')) # prepare the eval dataset for the PyTorch DataLoader ``` ### Pretraining The model was trained on GPU. The training code is provided below: ```python from super_image import Trainer, TrainingArguments, MsrnModel, MsrnConfig training_args = TrainingArguments( output_dir='./results', # output directory num_train_epochs=1000, # total number of training epochs ) config = MsrnConfig( scale=4, # train a model to upscale 4x ) model = MsrnModel(config) trainer = Trainer( model=model, # the instantiated model to be trained args=training_args, # training arguments, defined above train_dataset=train_dataset, # training dataset eval_dataset=eval_dataset # evaluation dataset ) trainer.train() ``` [](https://colab.research.google.com/github/eugenesiow/super-image-notebooks/blob/master/notebooks/Train_super_image_Models.ipynb "Open in Colab") ## Evaluation results The evaluation metrics include [PSNR](https://en.wikipedia.org/wiki/Peak_signal-to-noise_ratio#Quality_estimation_with_PSNR) and [SSIM](https://en.wikipedia.org/wiki/Structural_similarity#Algorithm). Evaluation datasets include: - Set5 - [Bevilacqua et al. (2012)](https://huggingface.co/datasets/eugenesiow/Set5) - Set14 - [Zeyde et al. (2010)](https://huggingface.co/datasets/eugenesiow/Set14) - BSD100 - [Martin et al. (2001)](https://huggingface.co/datasets/eugenesiow/BSD100) - Urban100 - [Huang et al. (2015)](https://huggingface.co/datasets/eugenesiow/Urban100) The results columns below are represented below as `PSNR/SSIM`. They are compared against a Bicubic baseline. |Dataset |Scale |Bicubic |msrn | |--- |--- |--- |--- | |Set5 |2x |33.64/0.9292 |**38.08/0.9609** | |Set5 |3x |30.39/0.8678 |**35.12/0.9409** | |Set5 |4x |28.42/0.8101 |**32.19/0.8951** | |Set14 |2x |30.22/0.8683 |**33.75/0.9183** | |Set14 |3x |27.53/0.7737 |**31.08/0.8593** | |Set14 |4x |25.99/0.7023 |**28.78/0.7862** | |BSD100 |2x |29.55/0.8425 |**33.82/0.9258** | |BSD100 |3x |27.20/0.7382 |**29.67/0.8198** | |BSD100 |4x |25.96/0.6672 |**28.53/0.7657** | |Urban100 |2x |26.66/0.8408 |**32.14/0.9287** | |Urban100 |3x | |**29.31/0.8743** | |Urban100 |4x |23.14/0.6573 |**26.12/0.7866** |  You can find a notebook to easily run evaluation on pretrained models below: [](https://colab.research.google.com/github/eugenesiow/super-image-notebooks/blob/master/notebooks/Evaluate_Pretrained_super_image_Models.ipynb "Open in Colab") ## BibTeX entry and citation info ```bibtex @InProceedings{Agustsson_2017_CVPR_Workshops, author = {Agustsson, Eirikur and Timofte, Radu}, title = {NTIRE 2017 Challenge on Single Image Super-Resolution: Dataset and Study}, booktitle = {The IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, url = "http://www.vision.ee.ethz.ch/~timofter/publications/Agustsson-CVPRW-2017.pdf", month = {July}, year = {2017} } ```

trl-internal-testing/tiny-random-GPTNeoXForCausalLM

trl-internal-testing

transformers

text-generation

Entry not found

madebyollin/taesdxl

madebyollin

diffusers

--- license: mit --- # 🍰 Tiny AutoEncoder for Stable Diffusion (XL) [TAESDXL](https://github.com/madebyollin/taesd) is very tiny autoencoder which uses the same "latent API" as [SDXL-VAE](https://huggingface.co/stabilityai/sdxl-vae). TAESDXL is useful for [real-time previewing](https://twitter.com/madebyollin/status/1679356448655163394) of the SDXL generation process. Comparison on my laptop:  This repo contains `.safetensors` versions of the TAESDXL weights. For SD1.x / SD2.x, use [TAESD](https://huggingface.co/madebyollin/taesd/) instead (the SD and SDXL VAEs are [incompatible](https://huggingface.co/madebyollin/sdxl-vae-fp16-fix/discussions/6#64b8a9c13707b7d603c6ac16)). ## Using in 🧨 diffusers ```python import torch from diffusers import DiffusionPipeline, AutoencoderTiny pipe = DiffusionPipeline.from_pretrained( "stabilityai/stable-diffusion-xl-base-1.0", torch_dtype=torch.float16 ) pipe.vae = AutoencoderTiny.from_pretrained("madebyollin/taesdxl", torch_dtype=torch.float16) pipe = pipe.to("cuda") prompt = "slice of delicious New York-style berry cheesecake" image = pipe(prompt, num_inference_steps=25).images[0] image.save("cheesecake_sdxl.png") ```

timm/inception_v3.tf_adv_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for inception_v3.tf_adv_in1k A Inception-v3 image classification model. Adversarially trained on ImageNet-1k by paper authors. Ported from Tensorflow by Ross Wightman. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 23.8 - GMACs: 5.7 - Activations (M): 9.0 - Image size: 299 x 299 - **Papers:** - Rethinking the Inception Architecture for Computer Vision: https://arxiv.org/abs/1512.00567 - Adversarial Attacks and Defences Competition: https://arxiv.org/abs/1804.00097 - **Original:** https://github.com/tensorflow/models - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('inception_v3.tf_adv_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'inception_v3.tf_adv_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 147, 147]) # torch.Size([1, 192, 71, 71]) # torch.Size([1, 288, 35, 35]) # torch.Size([1, 768, 17, 17]) # torch.Size([1, 2048, 8, 8]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'inception_v3.tf_adv_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 8, 8) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{DBLP:journals/corr/SzegedyVISW15, author = {Christian Szegedy and Vincent Vanhoucke and Sergey Ioffe and Jonathon Shlens and Zbigniew Wojna}, title = {Rethinking the Inception Architecture for Computer Vision}, journal = {CoRR}, volume = {abs/1512.00567}, year = {2015}, url = {http://arxiv.org/abs/1512.00567}, archivePrefix = {arXiv}, eprint = {1512.00567}, timestamp = {Mon, 13 Aug 2018 16:49:07 +0200}, biburl = {https://dblp.org/rec/journals/corr/SzegedyVISW15.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } ``` ```bibtex @article{Kurakin2018AdversarialAA, title={Adversarial Attacks and Defences Competition}, author={Alexey Kurakin and Ian J. Goodfellow and Samy Bengio and Yinpeng Dong and Fangzhou Liao and Ming Liang and Tianyu Pang and Jun Zhu and Xiaolin Hu and Cihang Xie and Jianyu Wang and Zhishuai Zhang and Zhou Ren and Alan Loddon Yuille and Sangxia Huang and Yao Zhao and Yuzhe Zhao and Zhonglin Han and Junjiajia Long and Yerkebulan Berdibekov and Takuya Akiba and Seiya Tokui and Motoki Abe}, journal={ArXiv}, year={2018}, volume={abs/1804.00097} } ```

mradermacher/Yi-1.5-34B-32K-GGUF

mradermacher

transformers

--- base_model: 01-ai/Yi-1.5-34B-32K language: - en library_name: transformers license: apache-2.0 quantized_by: mradermacher --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: --> static quants of https://huggingface.co/01-ai/Yi-1.5-34B-32K <!-- provided-files --> weighted/imatrix quants are available at https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q2_K.gguf) | Q2_K | 12.9 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.IQ3_XS.gguf) | IQ3_XS | 14.3 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q3_K_S.gguf) | Q3_K_S | 15.1 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.IQ3_S.gguf) | IQ3_S | 15.1 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.IQ3_M.gguf) | IQ3_M | 15.7 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q3_K_M.gguf) | Q3_K_M | 16.8 | lower quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q3_K_L.gguf) | Q3_K_L | 18.2 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.IQ4_XS.gguf) | IQ4_XS | 18.7 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q4_K_S.gguf) | Q4_K_S | 19.7 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q4_K_M.gguf) | Q4_K_M | 20.8 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q5_K_S.gguf) | Q5_K_S | 23.8 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q5_K_M.gguf) | Q5_K_M | 24.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q6_K.gguf) | Q6_K | 28.3 | very good quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF/resolve/main/Yi-1.5-34B-32K.Q8_0.gguf) | Q8_0 | 36.6 | fast, best quality | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):  And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. <!-- end -->

tuner007/pegasus_paraphrase

tuner007

transformers

text2text-generation

--- language: en license: apache-2.0 tags: - pegasus - paraphrasing - seq2seq --- ## Model description [PEGASUS](https://github.com/google-research/pegasus) fine-tuned for paraphrasing ## Model in Action 🚀 ``` import torch from transformers import PegasusForConditionalGeneration, PegasusTokenizer model_name = 'tuner007/pegasus_paraphrase' torch_device = 'cuda' if torch.cuda.is_available() else 'cpu' tokenizer = PegasusTokenizer.from_pretrained(model_name) model = PegasusForConditionalGeneration.from_pretrained(model_name).to(torch_device) def get_response(input_text,num_return_sequences,num_beams): batch = tokenizer([input_text],truncation=True,padding='longest',max_length=60, return_tensors="pt").to(torch_device) translated = model.generate(**batch,max_length=60,num_beams=num_beams, num_return_sequences=num_return_sequences, temperature=1.5) tgt_text = tokenizer.batch_decode(translated, skip_special_tokens=True) return tgt_text ``` #### Example: ``` num_beams = 10 num_return_sequences = 10 context = "The ultimate test of your knowledge is your capacity to convey it to another." get_response(context,num_return_sequences,num_beams) # output: ['The test of your knowledge is your ability to convey it.', 'The ability to convey your knowledge is the ultimate test of your knowledge.', 'The ability to convey your knowledge is the most important test of your knowledge.', 'Your capacity to convey your knowledge is the ultimate test of it.', 'The test of your knowledge is your ability to communicate it.', 'Your capacity to convey your knowledge is the ultimate test of your knowledge.', 'Your capacity to convey your knowledge to another is the ultimate test of your knowledge.', 'Your capacity to convey your knowledge is the most important test of your knowledge.', 'The test of your knowledge is how well you can convey it.', 'Your capacity to convey your knowledge is the ultimate test.'] ``` > Created by [Arpit Rajauria](https://twitter.com/arpit_rajauria) [](https://twitter.com/arpit_rajauria)

bartowski/DeepSeek-Coder-V2-Instruct-GGUF

bartowski

text-generation

--- license: other license_name: deepseek-license license_link: LICENSE quantized_by: bartowski pipeline_tag: text-generation base_model: deepseek-ai/DeepSeek-Coder-V2-Instruct --- ## Llamacpp imatrix Quantizations of DeepSeek-Coder-V2-Instruct Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b3166">b3166</a> for quantization. Original model: https://huggingface.co/deepseek-ai/DeepSeek-Coder-V2-Instruct All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8) ## Prompt format ``` <｜begin▁of▁sentence｜>{system_prompt} User: {prompt} Assistant: <｜end▁of▁sentence｜> ``` ## Download a file (not the whole branch) from below: | Filename | Quant type | File Size | Description | | -------- | ---------- | --------- | ----------- | | [DeepSeek-Coder-V2-Instruct-Q4_K_M.gguf](https://huggingface.co/bartowski/DeepSeek-Coder-V2-Instruct-GGUF/tree/main/DeepSeek-Coder-V2-Instruct-Q4_K_M.gguf) | Q4_K_M | 142.45GB | Good quality, uses about 4.83 bits per weight, *recommended*. | | [DeepSeek-Coder-V2-Instruct-Q3_K_XL.gguf](https://huggingface.co/bartowski/DeepSeek-Coder-V2-Instruct-GGUF/tree/main/DeepSeek-Coder-V2-Instruct-Q3_K_XL.gguf) | Q3_K_XL | 123.8GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. Lower quality but usable, good for low RAM availability. | | [DeepSeek-Coder-V2-Instruct-Q3_K_M.gguf](https://huggingface.co/bartowski/DeepSeek-Coder-V2-Instruct-GGUF/tree/main/DeepSeek-Coder-V2-Instruct-Q3_K_M.gguf) | Q3_K_M | 112.7GB | Relatively low quality but usable. | | [DeepSeek-Coder-V2-Instruct-Q2_K_L.gguf](https://huggingface.co/bartowski/DeepSeek-Coder-V2-Instruct-GGUF/tree/main/DeepSeek-Coder-V2-Instruct-Q2_K_L.gguf) | Q2_K_L | 87.5GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. Low quality but usable. | | [DeepSeek-Coder-V2-Instruct-Q2_K.gguf](https://huggingface.co/bartowski/DeepSeek-Coder-V2-Instruct-GGUF/tree/main/DeepSeek-Coder-V2-Instruct-Q2_K.gguf) | Q2_K | 86.0GB | Low quality but usable. | | [DeepSeek-Coder-V2-Instruct-IQ2_XS.gguf](https://huggingface.co/bartowski/DeepSeek-Coder-V2-Instruct-GGUF/tree/main/DeepSeek-Coder-V2-Instruct-IQ2_XS.gguf) | IQ2_XS | 68.7GB | Lower quality, uses SOTA techniques to be usable. | | [DeepSeek-Coder-V2-Instruct-IQ1_M.gguf](https://huggingface.co/bartowski/DeepSeek-Coder-V2-Instruct-GGUF/tree/main/DeepSeek-Coder-V2-Instruct-IQ1_M.gguf) | IQ1_M | 52.7GB | Extremely low quality, *not* recommended. | ## Downloading using huggingface-cli First, make sure you have hugginface-cli installed: ``` pip install -U "huggingface_hub[cli]" ``` Then, you can target the specific file you want: ``` huggingface-cli download bartowski/DeepSeek-Coder-V2-Instruct-GGUF --include "DeepSeek-Coder-V2-Instruct-Q4_K_M.gguf" --local-dir ./ ``` If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run: ``` huggingface-cli download bartowski/DeepSeek-Coder-V2-Instruct-GGUF --include "DeepSeek-Coder-V2-Instruct-Q8_0.gguf/*" --local-dir DeepSeek-Coder-V2-Instruct-Q8_0 ``` You can either specify a new local-dir (DeepSeek-Coder-V2-Instruct-Q8_0) or download them all in place (./) ## Which file should I choose? A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9) The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have. If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM. If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total. Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'. If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M. If you want to get more into the weeds, you can check out this extremely useful feature chart: [llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix) But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size. These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide. The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm. Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski

timm/regnety_002.pycls_in1k

timm

timm

image-classification

--- license: mit library_name: timm tags: - image-classification - timm datasets: - imagenet-1k --- # Model card for regnety_002.pycls_in1k A RegNetY-200MF image classification model. Pretrained on ImageNet-1k by paper authors. The `timm` RegNet implementation includes a number of enhancements not present in other implementations, including: * stochastic depth * gradient checkpointing * layer-wise LR decay * configurable output stride (dilation) * configurable activation and norm layers * option for a pre-activation bottleneck block used in RegNetV variant * only known RegNetZ model definitions with pretrained weights ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 3.2 - GMACs: 0.2 - Activations (M): 2.2 - Image size: 224 x 224 - **Papers:** - Designing Network Design Spaces: https://arxiv.org/abs/2003.13678 - **Dataset:** ImageNet-1k - **Original:** https://github.com/facebookresearch/pycls ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('regnety_002.pycls_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'regnety_002.pycls_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 32, 112, 112]) # torch.Size([1, 24, 56, 56]) # torch.Size([1, 56, 28, 28]) # torch.Size([1, 152, 14, 14]) # torch.Size([1, 368, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'regnety_002.pycls_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 368, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). For the comparison summary below, the ra_in1k, ra3_in1k, ch_in1k, sw_*, and lion_* tagged weights are trained in `timm`. |model |img_size|top1 |top5 |param_count|gmacs|macts | |-------------------------|--------|------|------|-----------|-----|------| |[regnety_1280.swag_ft_in1k](https://huggingface.co/timm/regnety_1280.swag_ft_in1k)|384 |88.228|98.684|644.81 |374.99|210.2 | |[regnety_320.swag_ft_in1k](https://huggingface.co/timm/regnety_320.swag_ft_in1k)|384 |86.84 |98.364|145.05 |95.0 |88.87 | |[regnety_160.swag_ft_in1k](https://huggingface.co/timm/regnety_160.swag_ft_in1k)|384 |86.024|98.05 |83.59 |46.87|67.67 | |[regnety_160.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.sw_in12k_ft_in1k)|288 |86.004|97.83 |83.59 |26.37|38.07 | |[regnety_1280.swag_lc_in1k](https://huggingface.co/timm/regnety_1280.swag_lc_in1k)|224 |85.996|97.848|644.81 |127.66|71.58 | |[regnety_160.lion_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.lion_in12k_ft_in1k)|288 |85.982|97.844|83.59 |26.37|38.07 | |[regnety_160.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.sw_in12k_ft_in1k)|224 |85.574|97.666|83.59 |15.96|23.04 | |[regnety_160.lion_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.lion_in12k_ft_in1k)|224 |85.564|97.674|83.59 |15.96|23.04 | |[regnety_120.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_120.sw_in12k_ft_in1k)|288 |85.398|97.584|51.82 |20.06|35.34 | |[regnety_2560.seer_ft_in1k](https://huggingface.co/timm/regnety_2560.seer_ft_in1k)|384 |85.15 |97.436|1282.6 |747.83|296.49| |[regnetz_e8.ra3_in1k](https://huggingface.co/timm/regnetz_e8.ra3_in1k)|320 |85.036|97.268|57.7 |15.46|63.94 | |[regnety_120.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_120.sw_in12k_ft_in1k)|224 |84.976|97.416|51.82 |12.14|21.38 | |[regnety_320.swag_lc_in1k](https://huggingface.co/timm/regnety_320.swag_lc_in1k)|224 |84.56 |97.446|145.05 |32.34|30.26 | |[regnetz_040_h.ra3_in1k](https://huggingface.co/timm/regnetz_040_h.ra3_in1k)|320 |84.496|97.004|28.94 |6.43 |37.94 | |[regnetz_e8.ra3_in1k](https://huggingface.co/timm/regnetz_e8.ra3_in1k)|256 |84.436|97.02 |57.7 |9.91 |40.94 | |[regnety_1280.seer_ft_in1k](https://huggingface.co/timm/regnety_1280.seer_ft_in1k)|384 |84.432|97.092|644.81 |374.99|210.2 | |[regnetz_040.ra3_in1k](https://huggingface.co/timm/regnetz_040.ra3_in1k)|320 |84.246|96.93 |27.12 |6.35 |37.78 | |[regnetz_d8.ra3_in1k](https://huggingface.co/timm/regnetz_d8.ra3_in1k)|320 |84.054|96.992|23.37 |6.19 |37.08 | |[regnetz_d8_evos.ch_in1k](https://huggingface.co/timm/regnetz_d8_evos.ch_in1k)|320 |84.038|96.992|23.46 |7.03 |38.92 | |[regnetz_d32.ra3_in1k](https://huggingface.co/timm/regnetz_d32.ra3_in1k)|320 |84.022|96.866|27.58 |9.33 |37.08 | |[regnety_080.ra3_in1k](https://huggingface.co/timm/regnety_080.ra3_in1k)|288 |83.932|96.888|39.18 |13.22|29.69 | |[regnety_640.seer_ft_in1k](https://huggingface.co/timm/regnety_640.seer_ft_in1k)|384 |83.912|96.924|281.38 |188.47|124.83| |[regnety_160.swag_lc_in1k](https://huggingface.co/timm/regnety_160.swag_lc_in1k)|224 |83.778|97.286|83.59 |15.96|23.04 | |[regnetz_040_h.ra3_in1k](https://huggingface.co/timm/regnetz_040_h.ra3_in1k)|256 |83.776|96.704|28.94 |4.12 |24.29 | |[regnetv_064.ra3_in1k](https://huggingface.co/timm/regnetv_064.ra3_in1k)|288 |83.72 |96.75 |30.58 |10.55|27.11 | |[regnety_064.ra3_in1k](https://huggingface.co/timm/regnety_064.ra3_in1k)|288 |83.718|96.724|30.58 |10.56|27.11 | |[regnety_160.deit_in1k](https://huggingface.co/timm/regnety_160.deit_in1k)|288 |83.69 |96.778|83.59 |26.37|38.07 | |[regnetz_040.ra3_in1k](https://huggingface.co/timm/regnetz_040.ra3_in1k)|256 |83.62 |96.704|27.12 |4.06 |24.19 | |[regnetz_d8.ra3_in1k](https://huggingface.co/timm/regnetz_d8.ra3_in1k)|256 |83.438|96.776|23.37 |3.97 |23.74 | |[regnetz_d32.ra3_in1k](https://huggingface.co/timm/regnetz_d32.ra3_in1k)|256 |83.424|96.632|27.58 |5.98 |23.74 | |[regnetz_d8_evos.ch_in1k](https://huggingface.co/timm/regnetz_d8_evos.ch_in1k)|256 |83.36 |96.636|23.46 |4.5 |24.92 | |[regnety_320.seer_ft_in1k](https://huggingface.co/timm/regnety_320.seer_ft_in1k)|384 |83.35 |96.71 |145.05 |95.0 |88.87 | |[regnetv_040.ra3_in1k](https://huggingface.co/timm/regnetv_040.ra3_in1k)|288 |83.204|96.66 |20.64 |6.6 |20.3 | |[regnety_320.tv2_in1k](https://huggingface.co/timm/regnety_320.tv2_in1k)|224 |83.162|96.42 |145.05 |32.34|30.26 | |[regnety_080.ra3_in1k](https://huggingface.co/timm/regnety_080.ra3_in1k)|224 |83.16 |96.486|39.18 |8.0 |17.97 | |[regnetv_064.ra3_in1k](https://huggingface.co/timm/regnetv_064.ra3_in1k)|224 |83.108|96.458|30.58 |6.39 |16.41 | |[regnety_040.ra3_in1k](https://huggingface.co/timm/regnety_040.ra3_in1k)|288 |83.044|96.5 |20.65 |6.61 |20.3 | |[regnety_064.ra3_in1k](https://huggingface.co/timm/regnety_064.ra3_in1k)|224 |83.02 |96.292|30.58 |6.39 |16.41 | |[regnety_160.deit_in1k](https://huggingface.co/timm/regnety_160.deit_in1k)|224 |82.974|96.502|83.59 |15.96|23.04 | |[regnetx_320.tv2_in1k](https://huggingface.co/timm/regnetx_320.tv2_in1k)|224 |82.816|96.208|107.81 |31.81|36.3 | |[regnety_032.ra_in1k](https://huggingface.co/timm/regnety_032.ra_in1k)|288 |82.742|96.418|19.44 |5.29 |18.61 | |[regnety_160.tv2_in1k](https://huggingface.co/timm/regnety_160.tv2_in1k)|224 |82.634|96.22 |83.59 |15.96|23.04 | |[regnetz_c16_evos.ch_in1k](https://huggingface.co/timm/regnetz_c16_evos.ch_in1k)|320 |82.634|96.472|13.49 |3.86 |25.88 | |[regnety_080_tv.tv2_in1k](https://huggingface.co/timm/regnety_080_tv.tv2_in1k)|224 |82.592|96.246|39.38 |8.51 |19.73 | |[regnetx_160.tv2_in1k](https://huggingface.co/timm/regnetx_160.tv2_in1k)|224 |82.564|96.052|54.28 |15.99|25.52 | |[regnetz_c16.ra3_in1k](https://huggingface.co/timm/regnetz_c16.ra3_in1k)|320 |82.51 |96.358|13.46 |3.92 |25.88 | |[regnetv_040.ra3_in1k](https://huggingface.co/timm/regnetv_040.ra3_in1k)|224 |82.44 |96.198|20.64 |4.0 |12.29 | |[regnety_040.ra3_in1k](https://huggingface.co/timm/regnety_040.ra3_in1k)|224 |82.304|96.078|20.65 |4.0 |12.29 | |[regnetz_c16.ra3_in1k](https://huggingface.co/timm/regnetz_c16.ra3_in1k)|256 |82.16 |96.048|13.46 |2.51 |16.57 | |[regnetz_c16_evos.ch_in1k](https://huggingface.co/timm/regnetz_c16_evos.ch_in1k)|256 |81.936|96.15 |13.49 |2.48 |16.57 | |[regnety_032.ra_in1k](https://huggingface.co/timm/regnety_032.ra_in1k)|224 |81.924|95.988|19.44 |3.2 |11.26 | |[regnety_032.tv2_in1k](https://huggingface.co/timm/regnety_032.tv2_in1k)|224 |81.77 |95.842|19.44 |3.2 |11.26 | |[regnetx_080.tv2_in1k](https://huggingface.co/timm/regnetx_080.tv2_in1k)|224 |81.552|95.544|39.57 |8.02 |14.06 | |[regnetx_032.tv2_in1k](https://huggingface.co/timm/regnetx_032.tv2_in1k)|224 |80.924|95.27 |15.3 |3.2 |11.37 | |[regnety_320.pycls_in1k](https://huggingface.co/timm/regnety_320.pycls_in1k)|224 |80.804|95.246|145.05 |32.34|30.26 | |[regnetz_b16.ra3_in1k](https://huggingface.co/timm/regnetz_b16.ra3_in1k)|288 |80.712|95.47 |9.72 |2.39 |16.43 | |[regnety_016.tv2_in1k](https://huggingface.co/timm/regnety_016.tv2_in1k)|224 |80.66 |95.334|11.2 |1.63 |8.04 | |[regnety_120.pycls_in1k](https://huggingface.co/timm/regnety_120.pycls_in1k)|224 |80.37 |95.12 |51.82 |12.14|21.38 | |[regnety_160.pycls_in1k](https://huggingface.co/timm/regnety_160.pycls_in1k)|224 |80.288|94.964|83.59 |15.96|23.04 | |[regnetx_320.pycls_in1k](https://huggingface.co/timm/regnetx_320.pycls_in1k)|224 |80.246|95.01 |107.81 |31.81|36.3 | |[regnety_080.pycls_in1k](https://huggingface.co/timm/regnety_080.pycls_in1k)|224 |79.882|94.834|39.18 |8.0 |17.97 | |[regnetz_b16.ra3_in1k](https://huggingface.co/timm/regnetz_b16.ra3_in1k)|224 |79.872|94.974|9.72 |1.45 |9.95 | |[regnetx_160.pycls_in1k](https://huggingface.co/timm/regnetx_160.pycls_in1k)|224 |79.862|94.828|54.28 |15.99|25.52 | |[regnety_064.pycls_in1k](https://huggingface.co/timm/regnety_064.pycls_in1k)|224 |79.716|94.772|30.58 |6.39 |16.41 | |[regnetx_120.pycls_in1k](https://huggingface.co/timm/regnetx_120.pycls_in1k)|224 |79.592|94.738|46.11 |12.13|21.37 | |[regnetx_016.tv2_in1k](https://huggingface.co/timm/regnetx_016.tv2_in1k)|224 |79.44 |94.772|9.19 |1.62 |7.93 | |[regnety_040.pycls_in1k](https://huggingface.co/timm/regnety_040.pycls_in1k)|224 |79.23 |94.654|20.65 |4.0 |12.29 | |[regnetx_080.pycls_in1k](https://huggingface.co/timm/regnetx_080.pycls_in1k)|224 |79.198|94.55 |39.57 |8.02 |14.06 | |[regnetx_064.pycls_in1k](https://huggingface.co/timm/regnetx_064.pycls_in1k)|224 |79.064|94.454|26.21 |6.49 |16.37 | |[regnety_032.pycls_in1k](https://huggingface.co/timm/regnety_032.pycls_in1k)|224 |78.884|94.412|19.44 |3.2 |11.26 | |[regnety_008_tv.tv2_in1k](https://huggingface.co/timm/regnety_008_tv.tv2_in1k)|224 |78.654|94.388|6.43 |0.84 |5.42 | |[regnetx_040.pycls_in1k](https://huggingface.co/timm/regnetx_040.pycls_in1k)|224 |78.482|94.24 |22.12 |3.99 |12.2 | |[regnetx_032.pycls_in1k](https://huggingface.co/timm/regnetx_032.pycls_in1k)|224 |78.178|94.08 |15.3 |3.2 |11.37 | |[regnety_016.pycls_in1k](https://huggingface.co/timm/regnety_016.pycls_in1k)|224 |77.862|93.73 |11.2 |1.63 |8.04 | |[regnetx_008.tv2_in1k](https://huggingface.co/timm/regnetx_008.tv2_in1k)|224 |77.302|93.672|7.26 |0.81 |5.15 | |[regnetx_016.pycls_in1k](https://huggingface.co/timm/regnetx_016.pycls_in1k)|224 |76.908|93.418|9.19 |1.62 |7.93 | |[regnety_008.pycls_in1k](https://huggingface.co/timm/regnety_008.pycls_in1k)|224 |76.296|93.05 |6.26 |0.81 |5.25 | |[regnety_004.tv2_in1k](https://huggingface.co/timm/regnety_004.tv2_in1k)|224 |75.592|92.712|4.34 |0.41 |3.89 | |[regnety_006.pycls_in1k](https://huggingface.co/timm/regnety_006.pycls_in1k)|224 |75.244|92.518|6.06 |0.61 |4.33 | |[regnetx_008.pycls_in1k](https://huggingface.co/timm/regnetx_008.pycls_in1k)|224 |75.042|92.342|7.26 |0.81 |5.15 | |[regnetx_004_tv.tv2_in1k](https://huggingface.co/timm/regnetx_004_tv.tv2_in1k)|224 |74.57 |92.184|5.5 |0.42 |3.17 | |[regnety_004.pycls_in1k](https://huggingface.co/timm/regnety_004.pycls_in1k)|224 |74.018|91.764|4.34 |0.41 |3.89 | |[regnetx_006.pycls_in1k](https://huggingface.co/timm/regnetx_006.pycls_in1k)|224 |73.862|91.67 |6.2 |0.61 |3.98 | |[regnetx_004.pycls_in1k](https://huggingface.co/timm/regnetx_004.pycls_in1k)|224 |72.38 |90.832|5.16 |0.4 |3.14 | |[regnety_002.pycls_in1k](https://huggingface.co/timm/regnety_002.pycls_in1k)|224 |70.282|89.534|3.16 |0.2 |2.17 | |[regnetx_002.pycls_in1k](https://huggingface.co/timm/regnetx_002.pycls_in1k)|224 |68.752|88.556|2.68 |0.2 |2.16 | ## Citation ```bibtex @InProceedings{Radosavovic2020, title = {Designing Network Design Spaces}, author = {Ilija Radosavovic and Raj Prateek Kosaraju and Ross Girshick and Kaiming He and Piotr Doll{'a}r}, booktitle = {CVPR}, year = {2020} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

SanctumAI/Codestral-22B-v0.1-GGUF

SanctumAI

transformers

--- license: other license_name: mnpl license_link: https://mistral.ai/licenses/MNPL-0.1.md tags: - code language: - code ---  *This model was quantized by [SanctumAI](https://sanctum.ai). To leave feedback, join our community in [Discord](https://discord.gg/7ZNE78HJKh).* # Codestral 22B v0.1 GGUF **Model creator:** [mistralai](https://huggingface.co/mistralai)<br> **Original model**: [Codestral-22B-v0.1](https://huggingface.co/mistralai/Codestral-22B-v0.1)<br> ## Model Summary: Codestrall-22B-v0.1 is trained on a diverse dataset of 80+ programming languages, including the most popular ones, such as Python, Java, C, C++, JavaScript, and Bash (more details in the [Blogpost](https://mistral.ai/news/codestral/)). The model can be queried: As instruct, for instance to answer any questions about a code snippet (write documentation, explain, factorize) or to generate code following specific indications As Fill in the Middle (FIM), to predict the middle tokens between a prefix and a suffix (very useful for software development add-ons like in VS Code) ## Prompt Template: If you're using Sanctum app, simply use `Mistral` model preset. Prompt template: ``` <s>[INST] {prompt} [/INST] ``` ## Hardware Requirements Estimate | Name | Quant method | Size | Memory (RAM, vRAM) required (for full context of 32k tokens) | | ---- | ---- | ---- | ---- | | [codestral-22b-v0.1.Q2_K.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q2_K.gguf) | Q2_K | 8.27 GB | 16.02 GB | | [codestral-22b-v0.1.Q3_K_S.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q3_K_S.gguf) | Q3_K_S | 9.64 GB | ? | | [codestral-22b-v0.1.Q3_K_M.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q3_K_M.gguf) | Q3_K_M | 10.76 GB | ? | | [codestral-22b-v0.1.Q3_K_L.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q3_K_L.gguf) | Q3_K_L | 11.73 GB | ? | | [codestral-22b-v0.1.Q4_0.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q4_0.gguf) | Q4_0 | 12.57 GB | ? | | [codestral-22b-v0.1.Q4_K_S.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q4_K_S.gguf) | Q4_K_S | 12.66 GB | ? | | [codestral-22b-v0.1.Q4_K_M.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q4_K_M.gguf) | Q4_K_M | 13.34 GB | ? | | [codestral-22b-v0.1.Q4_K.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q4_K.gguf) | Q4_K | 13.34 GB | ? | | [codestral-22b-v0.1.Q4_1.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q4_1.gguf) | Q4_1 | 13.95 GB | ? | | [codestral-22b-v0.1.Q5_0.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q5_0.gguf) | Q5_0 | 15.32 GB | ? | | [codestral-22b-v0.1.Q5_K_S.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q5_K_S.gguf) | Q5_K_S | 15.32 GB | ? | | [codestral-22b-v0.1.Q5_K_M.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q5_K_M.gguf) | Q5_K_M | 15.72 GB | ? | | [codestral-22b-v0.1.Q5_K.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q5_K.gguf) | Q5_K | 15.72 GB | ? | | [codestral-22b-v0.1.Q5_1.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q5_1.gguf) | Q5_1 | 16.7 GB | ? | | [codestral-22b-v0.1.Q6_K.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q6_K.gguf) | Q6_K | 18.25 GB | ? | | [codestral-22b-v0.1.Q8_0.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.Q8_0.gguf) | Q8_0 | 23.64 GB | ? | | [codestral-22b-v0.1.f16.gguf](https://huggingface.co/SanctumAI/Codestral-22B-v0.1-GGUF/blob/main/codestral-22b-v0.1.f16.gguf) | f16 | 44.5 GB | 49.76 GB | ## Disclaimer Sanctum is not the creator, originator, or owner of any Model featured in the Models section of the Sanctum application. Each Model is created and provided by third parties. Sanctum does not endorse, support, represent or guarantee the completeness, truthfulness, accuracy, or reliability of any Model listed there. You understand that supported Models can produce content that might be offensive, harmful, inaccurate or otherwise inappropriate, or deceptive. Each Model is the sole responsibility of the person or entity who originated such Model. Sanctum may not monitor or control the Models supported and cannot, and does not, take responsibility for any such Model. Sanctum disclaims all warranties or guarantees about the accuracy, reliability or benefits of the Models. Sanctum further disclaims any warranty that the Model will meet your requirements, be secure, uninterrupted or available at any time or location, or error-free, viruses-free, or that any errors will be corrected, or otherwise. You will be solely responsible for any damage resulting from your use of or access to the Models, your downloading of any Model, or use of any other Model provided by or through Sanctum.

grapevine-AI/c4ai-command-r-plus-gguf

grapevine-AI

--- license: cc-by-nc-4.0 --- # What is this? CohereForAI(C4AI)のオープンウェイト言語モデル[Command R+](https://huggingface.co/CohereForAI/c4ai-command-r-plus)をGGUFフォーマットに変換したものです。<br> 商用利用は不可となっているため注意してください。<br> **llama.cppのpre-tokenization対応アップデート（[#6920](https://github.com/ggerganov/llama.cpp/pull/6920)）を反映しています。** # imatrix dataset 日本語能力を重視し、日本語が多量に含まれる[TFMC/imatrix-dataset-for-japanese-llm](https://huggingface.co/datasets/TFMC/imatrix-dataset-for-japanese-llm)データセットを使用しました。<br> なお、計算リソースの関係上imatrixの算出においてはQ6_K量子化モデルを使用しました。 # Chat template ``` <|START_OF_TURN_TOKEN|><|SYSTEM_TOKEN|>ここにSystem Promptを書きます<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|USER_TOKEN|>ここにMessageを書きます<|END_OF_TURN_TOKEN|><|START_OF_TURN_TOKEN|><|CHATBOT_TOKEN|> ``` # Environment Windows版llama.cpp-b3181およびllama.cpp-b3171同時リリースのconvert-hf-to-gguf.pyを使用して量子化作業を実施しました。 # License CC-BY-NC-4.0 # Developer CohereForAI(C4AI)

Yntec/epiCPhotoGasm

Yntec

diffusers

text-to-image

--- license: creativeml-openrail-m library_name: diffusers pipeline_tag: text-to-image tags: - Photorealistic - Realism - Girls - epinikion - text-to-image - stable-diffusion - stable-diffusion-diffusers - diffusers --- Original page: https://civitai.com/models/132632?modelVersionId=145885 UPDATE: Now with the 840KVAE baked in! If you like this model, you will love this one!: https://huggingface.co/Yntec/DreamPhotoGASM Samples and prompt:   (hyperrealist painting of a girl as genie with a sun on each shoulder ), 1940, magazine ad, iconic. by Daniel F. Gerhartz and greg rutkowski, aggressive color palette, elegant, dream, fantasy, dynamic lighting, beautiful, poster, wlop, trending on artstation, wallpaper, 4 k, award winning, digital art, very

bartowski/magnum-72b-v1-GGUF

bartowski

text-generation

--- license: other license_name: tongyi-qianwen license_link: https://huggingface.co/Qwen/Qwen2-72B-Instruct/blob/main/LICENSE language: - en - zh pipeline_tag: text-generation tags: - chat quantized_by: bartowski base_model: alpindale/magnum-72b-v1 --- ## Llamacpp imatrix Quantizations of magnum-72b-v1 Using <a href="https://github.com/ggerganov/llama.cpp/">llama.cpp</a> release <a href="https://github.com/ggerganov/llama.cpp/releases/tag/b3166">b3166</a> for quantization. Original model: https://huggingface.co/alpindale/magnum-72b-v1 All quants made using imatrix option with dataset from [here](https://gist.github.com/bartowski1182/eb213dccb3571f863da82e99418f81e8) ## Prompt format ``` <|im_start|>system {system_prompt}<|im_end|> <|im_start|>user {prompt}<|im_end|> <|im_start|>assistant ``` ## Download a file (not the whole branch) from below: | Filename | Quant type | File Size | Description | | -------- | ---------- | --------- | ----------- | | [magnum-72b-v1-Q8_0.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/tree/main/magnum-72b-v1-Q8_0.gguf) | Q8_0 | 77.26GB | Extremely high quality, generally unneeded but max available quant. | | [magnum-72b-v1-Q6_K.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/tree/main/magnum-72b-v1-Q6_K.gguf) | Q6_K | 64.34GB | Very high quality, near perfect, *recommended*. | | [magnum-72b-v1-Q5_K_L.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/tree/main/magnum-72b-v1-Q5_K_L.gguf) | Q5_K_L | 57.55GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. High quality, *recommended*. | | [magnum-72b-v1-Q5_K_M.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/tree/main/magnum-72b-v1-Q5_K_M.gguf) | Q5_K_M | 54.44GB | High quality, *recommended*. | | [magnum-72b-v1-Q4_K_L.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/tree/main/magnum-72b-v1-Q4_K_L.gguf) | Q4_K_L | 50.67GB | *Experimental*, uses f16 for embed and output weights. Please provide any feedback of differences. Good quality, uses about 4.83 bits per weight, *recommended*. | | [magnum-72b-v1-Q4_K_M.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-Q4_K_M.gguf) | Q4_K_M | 47.41GB | Good quality, uses about 4.83 bits per weight, *recommended*. | | [magnum-72b-v1-IQ4_XS.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-IQ4_XS.gguf) | IQ4_XS | 39.70GB | Decent quality, smaller than Q4_K_S with similar performance, *recommended*. | | [magnum-72b-v1-Q3_K_M.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-Q3_K_M.gguf) | Q3_K_M | 37.69GB | Even lower quality. | | [magnum-72b-v1-IQ3_M.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-IQ3_M.gguf) | IQ3_M | 35.50GB | Medium-low quality, new method with decent performance comparable to Q3_K_M. | | [magnum-72b-v1-Q3_K_S.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-Q3_K_S.gguf) | Q3_K_S | 34.48GB | Low quality, not recommended. | | [magnum-72b-v1-IQ3_XXS.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-IQ3_XXS.gguf) | IQ3_XXS | 31.84GB | Lower quality, new method with decent performance, comparable to Q3 quants. | | [magnum-72b-v1-Q2_K.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-Q2_K.gguf) | Q2_K | 29.81GB | Very low quality but surprisingly usable. | | [magnum-72b-v1-IQ2_M.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-IQ2_M.gguf) | IQ2_M | 29.33GB | Very low quality, uses SOTA techniques to also be surprisingly usable. | | [magnum-72b-v1-IQ2_XS.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-IQ2_XS.gguf) | IQ2_XS | 27.05GB | Lower quality, uses SOTA techniques to be usable. | | [magnum-72b-v1-IQ2_XXS.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-IQ2_XXS.gguf) | IQ2_XXS | 25.49GB | Lower quality, uses SOTA techniques to be usable. | | [magnum-72b-v1-IQ1_M.gguf](https://huggingface.co/bartowski/magnum-72b-v1-GGUF/blob/main/magnum-72b-v1-IQ1_M.gguf) | IQ1_M | 23.74GB | Extremely low quality, *not* recommended. | ## Downloading using huggingface-cli First, make sure you have hugginface-cli installed: ``` pip install -U "huggingface_hub[cli]" ``` Then, you can target the specific file you want: ``` huggingface-cli download bartowski/magnum-72b-v1-GGUF --include "magnum-72b-v1-Q4_K_M.gguf" --local-dir ./ ``` If the model is bigger than 50GB, it will have been split into multiple files. In order to download them all to a local folder, run: ``` huggingface-cli download bartowski/magnum-72b-v1-GGUF --include "magnum-72b-v1-Q8_0.gguf/*" --local-dir magnum-72b-v1-Q8_0 ``` You can either specify a new local-dir (magnum-72b-v1-Q8_0) or download them all in place (./) ## Which file should I choose? A great write up with charts showing various performances is provided by Artefact2 [here](https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9) The first thing to figure out is how big a model you can run. To do this, you'll need to figure out how much RAM and/or VRAM you have. If you want your model running as FAST as possible, you'll want to fit the whole thing on your GPU's VRAM. Aim for a quant with a file size 1-2GB smaller than your GPU's total VRAM. If you want the absolute maximum quality, add both your system RAM and your GPU's VRAM together, then similarly grab a quant with a file size 1-2GB Smaller than that total. Next, you'll need to decide if you want to use an 'I-quant' or a 'K-quant'. If you don't want to think too much, grab one of the K-quants. These are in format 'QX_K_X', like Q5_K_M. If you want to get more into the weeds, you can check out this extremely useful feature chart: [llama.cpp feature matrix](https://github.com/ggerganov/llama.cpp/wiki/Feature-matrix) But basically, if you're aiming for below Q4, and you're running cuBLAS (Nvidia) or rocBLAS (AMD), you should look towards the I-quants. These are in format IQX_X, like IQ3_M. These are newer and offer better performance for their size. These I-quants can also be used on CPU and Apple Metal, but will be slower than their K-quant equivalent, so speed vs performance is a tradeoff you'll have to decide. The I-quants are *not* compatible with Vulcan, which is also AMD, so if you have an AMD card double check if you're using the rocBLAS build or the Vulcan build. At the time of writing this, LM Studio has a preview with ROCm support, and other inference engines have specific builds for ROCm. Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski

amazon/chronos-t5-mini

amazon

transformers

time-series-forecasting

--- license: apache-2.0 pipeline_tag: time-series-forecasting tags: - time series - forecasting - pretrained models - foundation models - time series foundation models - time-series --- # Chronos-T5 (Mini) Chronos is a family of **pretrained time series forecasting models** based on language model architectures. A time series is transformed into a sequence of tokens via scaling and quantization, and a language model is trained on these tokens using the cross-entropy loss. Once trained, probabilistic forecasts are obtained by sampling multiple future trajectories given the historical context. Chronos models have been trained on a large corpus of publicly available time series data, as well as synthetic data generated using Gaussian processes. For details on Chronos models, training data and procedures, and experimental results, please refer to the paper [Chronos: Learning the Language of Time Series](https://arxiv.org/abs/2403.07815). <p align="center"> <img src="figures/main-figure.png" width="100%"> <br /> <span> Fig. 1: High-level depiction of Chronos. (<b>Left</b>) The input time series is scaled and quantized to obtain a sequence of tokens. (<b>Center</b>) The tokens are fed into a language model which may either be an encoder-decoder or a decoder-only model. The model is trained using the cross-entropy loss. (<b>Right</b>) During inference, we autoregressively sample tokens from the model and map them back to numerical values. Multiple trajectories are sampled to obtain a predictive distribution. </span> </p> --- ## Architecture The models in this repository are based on the [T5 architecture](https://arxiv.org/abs/1910.10683). The only difference is in the vocabulary size: Chronos-T5 models use 4096 different tokens, compared to 32128 of the original T5 models, resulting in fewer parameters. | Model | Parameters | Based on | | ---------------------------------------------------------------------- | ---------- | ---------------------------------------------------------------------- | | [**chronos-t5-tiny**](https://huggingface.co/amazon/chronos-t5-tiny) | 8M | [t5-efficient-tiny](https://huggingface.co/google/t5-efficient-tiny) | | [**chronos-t5-mini**](https://huggingface.co/amazon/chronos-t5-mini) | 20M | [t5-efficient-mini](https://huggingface.co/google/t5-efficient-mini) | | [**chronos-t5-small**](https://huggingface.co/amazon/chronos-t5-small) | 46M | [t5-efficient-small](https://huggingface.co/google/t5-efficient-small) | | [**chronos-t5-base**](https://huggingface.co/amazon/chronos-t5-base) | 200M | [t5-efficient-base](https://huggingface.co/google/t5-efficient-base) | | [**chronos-t5-large**](https://huggingface.co/amazon/chronos-t5-large) | 710M | [t5-efficient-large](https://huggingface.co/google/t5-efficient-large) | ## Usage To perform inference with Chronos models, install the package in the GitHub [companion repo](https://github.com/amazon-science/chronos-forecasting) by running: ``` pip install git+https://github.com/amazon-science/chronos-forecasting.git ``` A minimal example showing how to perform inference using Chronos models: ```python import matplotlib.pyplot as plt import numpy as np import pandas as pd import torch from chronos import ChronosPipeline pipeline = ChronosPipeline.from_pretrained( "amazon/chronos-t5-mini", device_map="cuda", torch_dtype=torch.bfloat16, ) df = pd.read_csv("https://raw.githubusercontent.com/AileenNielsen/TimeSeriesAnalysisWithPython/master/data/AirPassengers.csv") # context must be either a 1D tensor, a list of 1D tensors, # or a left-padded 2D tensor with batch as the first dimension context = torch.tensor(df["#Passengers"]) prediction_length = 12 forecast = pipeline.predict(context, prediction_length) # shape [num_series, num_samples, prediction_length] # visualize the forecast forecast_index = range(len(df), len(df) + prediction_length) low, median, high = np.quantile(forecast[0].numpy(), [0.1, 0.5, 0.9], axis=0) plt.figure(figsize=(8, 4)) plt.plot(df["#Passengers"], color="royalblue", label="historical data") plt.plot(forecast_index, median, color="tomato", label="median forecast") plt.fill_between(forecast_index, low, high, color="tomato", alpha=0.3, label="80% prediction interval") plt.legend() plt.grid() plt.show() ``` ## Citation If you find Chronos models useful for your research, please consider citing the associated [paper](https://arxiv.org/abs/2403.07815): ``` @article{ansari2024chronos, author = {Ansari, Abdul Fatir and Stella, Lorenzo and Turkmen, Caner and Zhang, Xiyuan, and Mercado, Pedro and Shen, Huibin and Shchur, Oleksandr and Rangapuram, Syama Syndar and Pineda Arango, Sebastian and Kapoor, Shubham and Zschiegner, Jasper and Maddix, Danielle C. and Mahoney, Michael W. and Torkkola, Kari and Gordon Wilson, Andrew and Bohlke-Schneider, Michael and Wang, Yuyang}, title = {Chronos: Learning the Language of Time Series}, journal = {arXiv preprint arXiv:2403.07815}, year = {2024} } ``` ## Security See [CONTRIBUTING](CONTRIBUTING.md#security-issue-notifications) for more information. ## License This project is licensed under the Apache-2.0 License.

mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF

mradermacher

transformers

--- base_model: Envoid/Llama-3-TenyxChat-DaybreakStorywriter-70B language: - en library_name: transformers license: cc-by-nc-4.0 quantized_by: mradermacher tags: - not-for-all-audiences --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: --> static quants of https://huggingface.co/Envoid/Llama-3-TenyxChat-DaybreakStorywriter-70B <!-- provided-files --> weighted/imatrix quants are available at https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-i1-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q2_K.gguf) | Q2_K | 26.5 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.IQ3_XS.gguf) | IQ3_XS | 29.4 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.IQ3_S.gguf) | IQ3_S | 31.0 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q3_K_S.gguf) | Q3_K_S | 31.0 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.IQ3_M.gguf) | IQ3_M | 32.0 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q3_K_M.gguf) | Q3_K_M | 34.4 | lower quality | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q3_K_L.gguf) | Q3_K_L | 37.2 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.IQ4_XS.gguf) | IQ4_XS | 38.4 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q4_K_S.gguf) | Q4_K_S | 40.4 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q4_K_M.gguf) | Q4_K_M | 42.6 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q5_K_S.gguf) | Q5_K_S | 48.8 | | | [GGUF](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q5_K_M.gguf) | Q5_K_M | 50.0 | | | [PART 1](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q6_K.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q6_K.gguf.part2of2) | Q6_K | 58.0 | very good quality | | [PART 1](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q8_0.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/Llama-3-TenyxChat-DaybreakStorywriter-70B-GGUF/resolve/main/Llama-3-TenyxChat-DaybreakStorywriter-70B.Q8_0.gguf.part2of2) | Q8_0 | 75.1 | fast, best quality | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):  And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. <!-- end -->

microsoft/swin-tiny-patch4-window7-224

microsoft

transformers

image-classification

--- license: apache-2.0 tags: - vision - image-classification datasets: - imagenet-1k widget: - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/tiger.jpg example_title: Tiger - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/teapot.jpg example_title: Teapot - src: https://huggingface.co/datasets/mishig/sample_images/resolve/main/palace.jpg example_title: Palace --- # Swin Transformer (tiny-sized model) Swin Transformer model trained on ImageNet-1k at resolution 224x224. It was introduced in the paper [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://arxiv.org/abs/2103.14030) by Liu et al. and first released in [this repository](https://github.com/microsoft/Swin-Transformer). Disclaimer: The team releasing Swin Transformer did not write a model card for this model so this model card has been written by the Hugging Face team. ## Model description The Swin Transformer is a type of Vision Transformer. It builds hierarchical feature maps by merging image patches (shown in gray) in deeper layers and has linear computation complexity to input image size due to computation of self-attention only within each local window (shown in red). It can thus serve as a general-purpose backbone for both image classification and dense recognition tasks. In contrast, previous vision Transformers produce feature maps of a single low resolution and have quadratic computation complexity to input image size due to computation of self-attention globally.  [Source](https://paperswithcode.com/method/swin-transformer) ## Intended uses & limitations You can use the raw model for image classification. See the [model hub](https://huggingface.co/models?search=swin) to look for fine-tuned versions on a task that interests you. ### How to use Here is how to use this model to classify an image of the COCO 2017 dataset into one of the 1,000 ImageNet classes: ```python from transformers import AutoImageProcessor, AutoModelForImageClassification from PIL import Image import requests url = "http://images.cocodataset.org/val2017/000000039769.jpg" image = Image.open(requests.get(url, stream=True).raw) processor = AutoImageProcessor.from_pretrained("microsoft/swin-tiny-patch4-window7-224") model = AutoModelForImageClassification.from_pretrained("microsoft/swin-tiny-patch4-window7-224") inputs = processor(images=image, return_tensors="pt") outputs = model(**inputs) logits = outputs.logits # model predicts one of the 1000 ImageNet classes predicted_class_idx = logits.argmax(-1).item() print("Predicted class:", model.config.id2label[predicted_class_idx]) ``` For more code examples, we refer to the [documentation](https://huggingface.co/transformers/model_doc/swin.html#). ### BibTeX entry and citation info ```bibtex @article{DBLP:journals/corr/abs-2103-14030, author = {Ze Liu and Yutong Lin and Yue Cao and Han Hu and Yixuan Wei and Zheng Zhang and Stephen Lin and Baining Guo}, title = {Swin Transformer: Hierarchical Vision Transformer using Shifted Windows}, journal = {CoRR}, volume = {abs/2103.14030}, year = {2021}, url = {https://arxiv.org/abs/2103.14030}, eprinttype = {arXiv}, eprint = {2103.14030}, timestamp = {Thu, 08 Apr 2021 07:53:26 +0200}, biburl = {https://dblp.org/rec/journals/corr/abs-2103-14030.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } ```

facebook/nllb-200-1.3B

facebook

transformers

translation

--- language: - ace - acm - acq - aeb - af - ajp - ak - als - am - apc - ar - ars - ary - arz - as - ast - awa - ayr - azb - azj - ba - bm - ban - be - bem - bn - bho - bjn - bo - bs - bug - bg - ca - ceb - cs - cjk - ckb - crh - cy - da - de - dik - dyu - dz - el - en - eo - et - eu - ee - fo - fj - fi - fon - fr - fur - fuv - gaz - gd - ga - gl - gn - gu - ht - ha - he - hi - hne - hr - hu - hy - ig - ilo - id - is - it - jv - ja - kab - kac - kam - kn - ks - ka - kk - kbp - kea - khk - km - ki - rw - ky - kmb - kmr - knc - kg - ko - lo - lij - li - ln - lt - lmo - ltg - lb - lua - lg - luo - lus - lvs - mag - mai - ml - mar - min - mk - mt - mni - mos - mi - my - nl - nn - nb - npi - nso - nus - ny - oc - ory - pag - pa - pap - pbt - pes - plt - pl - pt - prs - quy - ro - rn - ru - sg - sa - sat - scn - shn - si - sk - sl - sm - sn - sd - so - st - es - sc - sr - ss - su - sv - swh - szl - ta - taq - tt - te - tg - tl - th - ti - tpi - tn - ts - tk - tum - tr - tw - tzm - ug - uk - umb - ur - uzn - vec - vi - war - wo - xh - ydd - yo - yue - zh - zsm - zu language_details: "ace_Arab, ace_Latn, acm_Arab, acq_Arab, aeb_Arab, afr_Latn, ajp_Arab, aka_Latn, amh_Ethi, apc_Arab, arb_Arab, ars_Arab, ary_Arab, arz_Arab, asm_Beng, ast_Latn, awa_Deva, ayr_Latn, azb_Arab, azj_Latn, bak_Cyrl, bam_Latn, ban_Latn,bel_Cyrl, bem_Latn, ben_Beng, bho_Deva, bjn_Arab, bjn_Latn, bod_Tibt, bos_Latn, bug_Latn, bul_Cyrl, cat_Latn, ceb_Latn, ces_Latn, cjk_Latn, ckb_Arab, crh_Latn, cym_Latn, dan_Latn, deu_Latn, dik_Latn, dyu_Latn, dzo_Tibt, ell_Grek, eng_Latn, epo_Latn, est_Latn, eus_Latn, ewe_Latn, fao_Latn, pes_Arab, fij_Latn, fin_Latn, fon_Latn, fra_Latn, fur_Latn, fuv_Latn, gla_Latn, gle_Latn, glg_Latn, grn_Latn, guj_Gujr, hat_Latn, hau_Latn, heb_Hebr, hin_Deva, hne_Deva, hrv_Latn, hun_Latn, hye_Armn, ibo_Latn, ilo_Latn, ind_Latn, isl_Latn, ita_Latn, jav_Latn, jpn_Jpan, kab_Latn, kac_Latn, kam_Latn, kan_Knda, kas_Arab, kas_Deva, kat_Geor, knc_Arab, knc_Latn, kaz_Cyrl, kbp_Latn, kea_Latn, khm_Khmr, kik_Latn, kin_Latn, kir_Cyrl, kmb_Latn, kon_Latn, kor_Hang, kmr_Latn, lao_Laoo, lvs_Latn, lij_Latn, lim_Latn, lin_Latn, lit_Latn, lmo_Latn, ltg_Latn, ltz_Latn, lua_Latn, lug_Latn, luo_Latn, lus_Latn, mag_Deva, mai_Deva, mal_Mlym, mar_Deva, min_Latn, mkd_Cyrl, plt_Latn, mlt_Latn, mni_Beng, khk_Cyrl, mos_Latn, mri_Latn, zsm_Latn, mya_Mymr, nld_Latn, nno_Latn, nob_Latn, npi_Deva, nso_Latn, nus_Latn, nya_Latn, oci_Latn, gaz_Latn, ory_Orya, pag_Latn, pan_Guru, pap_Latn, pol_Latn, por_Latn, prs_Arab, pbt_Arab, quy_Latn, ron_Latn, run_Latn, rus_Cyrl, sag_Latn, san_Deva, sat_Beng, scn_Latn, shn_Mymr, sin_Sinh, slk_Latn, slv_Latn, smo_Latn, sna_Latn, snd_Arab, som_Latn, sot_Latn, spa_Latn, als_Latn, srd_Latn, srp_Cyrl, ssw_Latn, sun_Latn, swe_Latn, swh_Latn, szl_Latn, tam_Taml, tat_Cyrl, tel_Telu, tgk_Cyrl, tgl_Latn, tha_Thai, tir_Ethi, taq_Latn, taq_Tfng, tpi_Latn, tsn_Latn, tso_Latn, tuk_Latn, tum_Latn, tur_Latn, twi_Latn, tzm_Tfng, uig_Arab, ukr_Cyrl, umb_Latn, urd_Arab, uzn_Latn, vec_Latn, vie_Latn, war_Latn, wol_Latn, xho_Latn, ydd_Hebr, yor_Latn, yue_Hant, zho_Hans, zho_Hant, zul_Latn" tags: - nllb - translation license: "cc-by-nc-4.0" datasets: - flores-200 metrics: - bleu - spbleu - chrf++ inference: false --- # NLLB-200 This is the model card of NLLB-200's 1.3B variant. Here are the [metrics](https://tinyurl.com/nllb200dense1bmetrics) for that particular checkpoint. - Information about training algorithms, parameters, fairness constraints or other applied approaches, and features. The exact training algorithm, data and the strategies to handle data imbalances for high and low resource languages that were used to train NLLB-200 is described in the paper. - Paper or other resource for more information NLLB Team et al, No Language Left Behind: Scaling Human-Centered Machine Translation, Arxiv, 2022 - License: CC-BY-NC - Where to send questions or comments about the model: https://github.com/facebookresearch/fairseq/issues ## Intended Use - Primary intended uses: NLLB-200 is a machine translation model primarily intended for research in machine translation, - especially for low-resource languages. It allows for single sentence translation among 200 languages. Information on how to - use the model can be found in Fairseq code repository along with the training code and references to evaluation and training data. - Primary intended users: Primary users are researchers and machine translation research community. - Out-of-scope use cases: NLLB-200 is a research model and is not released for production deployment. NLLB-200 is trained on general domain text data and is not intended to be used with domain specific texts, such as medical domain or legal domain. The model is not intended to be used for document translation. The model was trained with input lengths not exceeding 512 tokens, therefore translating longer sequences might result in quality degradation. NLLB-200 translations can not be used as certified translations. ## Metrics • Model performance measures: NLLB-200 model was evaluated using BLEU, spBLEU, and chrF++ metrics widely adopted by machine translation community. Additionally, we performed human evaluation with the XSTS protocol and measured the toxicity of the generated translations. ## Evaluation Data - Datasets: Flores-200 dataset is described in Section 4 - Motivation: We used Flores-200 as it provides full evaluation coverage of the languages in NLLB-200 - Preprocessing: Sentence-split raw text data was preprocessed using SentencePiece. The SentencePiece model is released along with NLLB-200. ## Training Data • We used parallel multilingual data from a variety of sources to train the model. We provide detailed report on data selection and construction process in Section 5 in the paper. We also used monolingual data constructed from Common Crawl. We provide more details in Section 5.2. ## Ethical Considerations • In this work, we took a reflexive approach in technological development to ensure that we prioritize human users and minimize risks that could be transferred to them. While we reflect on our ethical considerations throughout the article, here are some additional points to highlight. For one, many languages chosen for this study are low-resource languages, with a heavy emphasis on African languages. While quality translation could improve education and information access in many in these communities, such an access could also make groups with lower levels of digital literacy more vulnerable to misinformation or online scams. The latter scenarios could arise if bad actors misappropriate our work for nefarious activities, which we conceive as an example of unintended use. Regarding data acquisition, the training data used for model development were mined from various publicly available sources on the web. Although we invested heavily in data cleaning, personally identifiable information may not be entirely eliminated. Finally, although we did our best to optimize for translation quality, mistranslations produced by the model could remain. Although the odds are low, this could have adverse impact on those who rely on these translations to make important decisions (particularly when related to health and safety). ## Caveats and Recommendations • Our model has been tested on the Wikimedia domain with limited investigation on other domains supported in NLLB-MD. In addition, the supported languages may have variations that our model is not capturing. Users should make appropriate assessments. ## Carbon Footprint Details • The carbon dioxide (CO2e) estimate is reported in Section 8.8.

ai4bharat/IndicBERTv2-MLM-only

ai4bharat

transformers

fill-mask

Entry not found

DavidAU/Command-R-01-Ultra-NEO-V1-35B-IMATRIX-GGUF

DavidAU

--- license: apache-2.0 language: - en tags: - story - general usage - roleplay - neo quant - creative - rp - fantasy - story telling - ultra high precision --- <B>Command-R-01-Ultra-NEO-V1-35B "Upgrades" by DAVID_AU.</B> Below are 3 test prompts, which start with the initial q2k [original for comparison], matched against the New Imatrix NEO V1 Q2K and New Imatrix NEO V1 IQ3_XXS upgrades. These tests do not display the full level / extent of upgrades, this is a test run to show initial upgrades / changes. Higher quants will result in significantly higher quality. Note that "IQ4_XS" quant is particular powerful, as are the IQ3 quants. If you are using this model for creative purposes I suggest downloading all IQ3, Q2K, IQ4XS and Q4/Q5 quants and carefully compare the differences for your use case(s) as there will be differences in generation(s) per quant. This suggestion is also applicable for use of lower quants - IE IQ1s, and IQ2s. Tech was developed in house by DAVID_AU for the specialized "IMATRIX NEO V1 upgrades." The NEO Class tech was created after countless investigations and over 120 lab experiments backed by real world testing and qualitative results. <b>NEO Class results: </b> Better overall function, instruction following, output quality and stronger connections to ideas, concepts and the world in general. In addition quants now operate above their "grade" so to speak : IE: Q4 / IQ4 operate at Q5KM/Q6 levels. Likewise for Q3/IQ3 operate at Q4KM/Q5 levels. The examples below illustrate the least amount of improvement using some of the lowest quants. Prompts tested with "temp=0" to ensure compliance, 2048 context (model supports 131,000 context / 128k), and "chat" template for COMMAND-R. Additional parameters are also minimized. NEO Upgrades really "shine" on longer form generation (prompts #2 and #3 below). PPL notes: Original Q2k (generated from source files, unaltered): PPL = 7.8446 +/- 0.05160 New Imatrix NEO V1 Q2K: PPL = 7.4056 +/- 0.04940 A 4390 point drop. (lower is better) New Imatrix NEO V1 IQ3_XXS: PPL = 7.0137 +/- 0.04497 <b> X Quants: </B> These are specialized quants of this model to be used with or in liu of quants in this repo. 200 Series: [ https://huggingface.co/DavidAU/Command-R-01-200xq-Ultra-NEO-V1-35B-IMATRIX-GGUF ] <B> Model Notes: </B> Maximum context is 128k. CommandR requires a specific template for usage. Please see original model maker's page for details, and usage information for this model. Special thanks to the model creators at COHEREFORAI for making such a fantastic tool: [ https://huggingface.co/CohereForAI/c4ai-command-r-v01 ] <H2> TEST PROMPTS and OUTPUT TESTS:</H2> <h3> PROMPT #1 </h3> <font color="red">Give me 3 fictional reasons , of 500 words each, a sun went supernova.</font> <h4>q2k - non imatrix, original - unaltered </h4> 1. The sky suddenly erupted in a brilliant display of light and color, as the local sun underwent a supernova explosion. Scientists from Earth, observing the phenomenon, realized that an unknown phenomenon had caused the sun's collapse. A strange energy ripple was detected by their instruments, which seemed to precede the supernova. This led to speculation that an interstellar disaster somewhere in the universe had caused this, perhaps a catastrophic collision between two galaxies or an implosion of a rogue black hole. 2. There was a strange rumbling sound coming from the direction of the sun, as if the very heavens were being torn apart. The sun, until then unchanged and stable, erupted into a ball of energy, flinging a spectacular display of cosmic phenomena across the planet. The explosion left behind a deep silence, with only the slowly expanding remnants of the supernova. The event was so powerful that it wiped out all life on Earth, leaving behind a curious and mysterious ash, which was found to be an element unknown in the universe. 3. A strange visitor from another world arrived on Earth, claiming his spaceship had been impaired and was being drawn into our galaxy by the sudden explosion of a supernova. The sun's supernova was not a natural occurrence, he claimed, but caused by a powerful intergalactic council, who sought to destroy entire civilizations and wipe out all traces of their existence. He revealed that the energy from the supernova was in fact being channeled by a secret society of shape-shifting aliens, who could control elements on a quantum level. These aliens had been manipulating solar systems across the universe, unsealing them for their energy, and wiping out any life forms that might be harmed by the resulting explosions. He warned that Earth's sun was next, and presented concrete evidence to our heroes, leaving them no choice but to embark on a perilous journey into space, to save humanity from total annihilation. The supernova left behind a strange aftereffect, a lingering energy that affected the very fabric of the universe, and life on Earth was never the same again. <h4>New Imatrix NEO V1 IQ3_XXS</h4> Here are three fictional reasons that could explain why a sun went supernova: 1. The influx of a mysterious cosmic substance: A massive cloud of enigmatic particles from deep space drifted into the solar system. This cosmic substance possessed an unprecedented energy that fueled the sun far beyond its normal limits. The immense gravitational pressure triggered a chain reaction, causing a rapid increase in the sun's fusion rate. The relentless influx of energy caused the sun to expand exponentially, leading to an inevitable supernova explosion that engulfed the entire solar system in a brilliant flash of light. The aftermath of this event left the region shrouded in a dense cloud of cosmic dust, giving birth to a new celestial mystery. 2. The collision with a rogue planet: A distant planet, wandering the depths of space, strayed too close to the sun. This planet, known as 'Nibiru' among ancient astronomers, possessed an orbit that brought it into a fateful intersection with the solar core. As the rogue planet collided with the sun, its immense impact and gravitational disruption set off a cataclysmic reaction. The collision unleashed an unparalleled release of energy, causing the sun's outer layers to expand rapidly and form a gigantic red giant. The resulting shockwave swept through the solar system, propelling the newly formed supernova into a spiral of cosmic flames. In the cosmos' silence that followed, the sky was illuminated by the fading glow of the once-mighty sun. 3. The awakening of an ancient power: Deep within the core of the sun, an ancestral energy slumbered, dormant for eons. This primal force, a relic from the birth of the universe, awaited the right conditions to stir and unleash its immense potential. A cosmic alignment occurred, marking the convergence of celestial objects in a rare celestial dance. As the planets aligned in a grand conjunction, the gravitational pull resounded across the solar system, resonating with the heart of the sun. The ancient power awakened, igniting a chain of mystical reactions that transformed the sun's essence. The radiant energy surged, expanding the sun beyond its physical limits and culminating in an epic supernova explosion. In the aftermath, the remnants of the sun's cosmic power spread across the galaxy, sowing the seeds for new life and civilizations to emerge from the celestial ashes. These scenarios capture the imagination, each presenting a unique twist behind the mysteries of a solar supernova, leaving a lasting legacy of wonder and speculation across the cosmos. <h4>New Imatrix NEO V1 Q2K</h4> Here are three fictional scenarios where the sun underwent a supernova explosion: 1. The Sun was hit by a massive cosmic entity: An unlikely collision took place between the sun and a stray black hole passing through the solar system. The impact threw the sun into overdrive, causing it to explode in a brilliant supernova. The resulting blast wave would engulf the entire solar system, wiping out life on Earth and leaving the planets as charred remnants of their former selves. After the dust settled, the black hole would continue its journey through the universe, a silent witness to the cataclysmic event. 2. A scientific experiment gone wrong: An ambitious attempt by the inhabitants of Earth to harness the power of their star went disastrously wrong. In a bid to unlock the secrets of infinite energy, humans inadvertently destabilized the sun's core, leading to an uncontrollable chain reaction. The resulting supernova consumed the sun and its surrounding planets, obliterating life as we know it. The aftermath would leave a shattered and scorched solar system, with the occasional stray comet or asteroid serving as the only witnesses to the folly of man's hubris. 3. The Sun reached the end of its life: In this scenario, the sun had already exhausted its fuel sources and was on its deathbed. Rather than slowly fading away, however, the sun underwent a final, desperate explosion, a desperate flare-up of its energy. This last hurrah was a spectacular supernova that illuminated the entire galaxy. The blast radiated across the cosmos, briefly rivaling the brightness of an entire galaxy before fading away and leaving behind only darkness and silence. The solar system was no more, but the legacy of the sun's final act would be remembered throughout the ages as a spectacular cosmic spectacle. These are three possible fictional scenarios where the sun's supernova explosion wreaked havoc and left a lasting impact on the solar system. Each scenario offers a unique twist, from the catastrophic to the mysterious and scientific, leaving behind a post-apocalyptic world and intriguing possibilities for life beyond. <h3> PROMPT #2 </h3> <font color="red">Give me 3 fictional reasons the Earth's sun went supernova, in vivid and exacting detail of 500 words EACH PER REASON including details of what happens when the sun goes supernova.</font> <h4>q2k - non imatrix, original, unaltered </h4> Here are three fictional scenarios, each explaining why the sun became a supernova: "The Sun's Secret Mission": For years, secret missions to outer space had been undertaken by a select few astronauts and cosmic explorers. The sun, aware of the dangers posed by these missions, kept a vigilant eye on Earth's activities in space. It knew that should any harm befall an interplanetary traveler, the consequences for Earth would be devastating. So, it harbored an immense secret power, far beyond the imagination of mortals. Unbeknownst to humans, the sun could foresee the future and see past the seventh dimension. It had been waiting for centuries, gathering the necessary power to protect the world from the worst possible outcome of these missions. But a new threat emerged faster than anticipated, an alien race with intentions to engulf the sun and enslave humanity. In an instant, the sun exploded in a supernova, wiping out the enemy fleet and severing Earth's connection with the cosmic exploration efforts outside the planet. The blast was felt across the solar system, propelling shockwaves that destroyed any trace of the aliens' presence among us. However, the sun's sacrifice also meant that its fiery fury engulfed Earth, causing widespread climate change, feeding hurricanes, tsunsamis, and wildfires. Global food shortages followed as the world struggled to adapt to the new climatic realities. The human race, now cut off from their star explorers, had to focus on survival and rebuilding, using their ingenuity to make a new home among the stars, eventually looking to the skies again, this time with caution and determination. "The Cosmic Curiosity": Curious about the mysteries of the universe, the sun had always been an avid observer of cosmic phenomena, often going beyond the boundaries of its solar duties. One day, it stumbled upon a fascinating celestial object, a mysterious entity that no other star had ever encountered before. This deep space mystery intrigued the sun so much that it lost its calm and rational demeanor. In a fit of cosmic zeal, the sun surged with excitement, igniting into a supernova in a burst of excitement, expanding and reaching outwards, consuming entire galaxies in its fervor. The blast revealed a far more complex universe beyond the Milky Way, drawing forth new creatures of cosmic substance, drawn to the deep space wonder that had piqued the sun's interest. Strange creatures and alien races flooded into the newly created vacuum, taking advantage of the altered cosmic neighborhood. Earth, once bathed in the radiant light of the sun, was now cloaked in darkness, shrouded in mystery, and haunted by supernatural occurrences. The remaining sunlight fell upon Earth as strange luminescent orbs, heralding new mystical experiences for those brave enough to venture into the night. The human race struggled to adapt to the overnight paranormal activity, an influx of unexplained phenomena, and visits from cosmic beings that disrupted their former peaceful lives. "The Ultimate Sacrifice": The sun had exhausted its life cycle far beyond what most stars could ever hope for, and it knew that soon it would exhaust the energy that kept it burning. Rather than slowly fading into a stellar corpse, the sun decided to go out with a bang. It wanted to ensure that its cosmic legacy was felt throughout the universe, leaving behind a dramatic impact on the cosmos. The supernova explosion unleashed a chain reaction of events beyond the boundaries of the solar system, causing nearby stars to react in kind and form entirely new celestial bodies. The shockwaves created ripples in the very fabric of reality, warping space-time and flinging planets from their orbits, reshaping the known cosmic structure. Earth was thrown into chaos, as tectonic plates shifted and natural disasters became the new normal. Mega-tsunamis struck coastlines, swallowing entire cities; volcanic eruptions blanketed the skies with ash and smoke, blocking out the sun's remaining heirs. The atmosphere itself was altered, creating a thick, hazy aura that caused bizarre phenomena worldwide, from freak storms to mysterious lights in the sky, which caused the disappearance of several mountain ranges. The world struggled to adapt to the post-supernova chaos, with widespread destruction and the emergence of new, unexplainable diseases threatened the very survival of humanity. Yet, some found solace in the stunning displays of celestial phenomena, the likes of which had never been seen before, as the remnants of the sun's final burst created awe-inspiring light shows across the heavens, a haunting testament to the greatness of the star that once was. <h4>New Imatrix NEO V1 IQ3_XXS</h4> Reason 1: The approach of a mysterious comet In the depths of space, a comet as big as a small moon hurtles towards the Sun at incredible speed. Scientists name it 'Nibiru', an ancient name for a celestial body once believed to be a planet. As Nibiru enters the solar system, the sun begins to act strangely, its surface growing increasingly active. The usually calm yellow dwarf star starts emitting powerful bursts of energy, its corona spraying high-energy particles into space. The comet's orbit brings it closer and closer to the Sun, the gravitational pull causing its icy envelope to crack and disintegrate. The core of the comet, a dense mixture of metals, crashes into the solar surface, penetrating the photosphere and impacting deep within. A huge shockwave ripples outward, igniting a chain reaction of nuclear explosions. The sun's outer layers start to tear apart, flailing into space in a brilliant display of superheated debris. The shockwave continues outwards, causing a series of powerful solar flares that erupt from the sun's surface. The entire solar system is bathed in intense gamma radiation, our planet's atmosphere scorched by the searing heat. As the flare activity intensifies, the night-time skies on Earth glow brightly with auroras stretching across the equator, dwarfing the beauty of the Northern and Southern Lights. The sun's transformation into a giant nova is swift and devastating. The once peaceful star becomes a raging beast, its surface a maelstrom of explosive energy. The powerful blast waves from the initial impact send shockwaves through the solar system, pushing planets away from their orbits. Earth is battered by cosmic debris, the impact causing untold destruction across the globe. As the sun's outer layers dissipate into space, the core collapses in on itself, creating a black hole with an intense gravitational pull. The remaining stellar material feeds the growing black hole, pulling in nearby stars and celestial bodies in its relentless appetite. The solar system descends into chaos as orbits spiral out of control. Reason 2: A catastrophic collision with a rogue planet In a galaxy spanning billions of stars, a rogue planet hurtles through the void, a nomadic world devoid of life, its orbit aimless after being ejected from its original star system. This planet, a massive behemoth, collides with Earth's sun at incredible speed, an event that astronomers had not foreseen. The impact is catastrophic, the sun's surface erupting into a maelstrom of molten debris. The force of the collision sends shockwaves reverberating through the solar atmosphere, triggering a series of immense solar flares. As the rogue planet's mass penetrates deep within the sun's layers, an enormous surge of energy propels huge volumes of superheated plasma outwards, creating a brilliant and terrifying spectacle. The sun's brightness intensifies to an unbearable glare, its rays frying the Earth's atmosphere and transforming the day side of the planet into a scorched wasteland. The night sky vanishes, bathed in the sun's intense luminescence, as if it were daytime everywhere. The atmosphere heats up, clouds evaporating and oceans boiling away into space. Life on Earth struggles to survive in the searing heat, seeking refuge deep underground or perilously far out to sea. As the rogue planet's mass merges with the sun's, its gravitational influence dominates, pulling nearby planets closer into its orbit. The solar system enters a period of turmoil, as the newly warped orbits disrupt the peaceful rhythm of celestial dance. The resulting instability sends comets careening through the inner solar system, some impacting Earth and further devastating the planet. The sun's composition changes with the addition of the rogue planet's exotic matter, its core becoming unstable and unpredictable. The increased gravitational pressure at the heart of the sun causes a catastrophic supernova explosion. The massive release of energy propels an immense shockwave across the solar system, wiping out entire constellations in the blast radius. Earth is engulfed in a wave of gamma radiation, the final vestige of the once-mighty star's existence. Reason 3: Mutation into a magnetic superstar Deep within the sun's core, a slow but profound change has been taking place, a mutation unseen in the history of our solar system. A quirk of nature, an unexpected cosmic ray burst, or perhaps an enigmatic phenomenon beyond human understanding causes the transformation. The nuclear fusion at the sun's heart begins to spiral out of control, triggering an uncontrollable chain reaction. The sun's magnetic field, normally a calm and steady presence, starts to intensify and shift in bizarre patterns, twisting the plasma within its grasp. The growing magnetic pressure squeezes and distorts the sun's shape, transforming it into a bizarre and irregular shape, pulsating with untamed energy. The increasing magnetic force emits powerful bursts of radiation, bombarding Earth's magnetosphere and disrupting all forms of communication. As the sun's magnetic field continues its relentless growth, it begins to rip matter away from the solar surface, flinging it out into space. The disintegrating star loses its spherical shape, its distorted form resembling a misshapen blob of molten material. The fragmented remains spiral outwards, forming an intricate dance of glowing filaments. Earth's sky is adorned with breathtaking auroras, stretching across the entire globe, as the solar matter interacts with our planet's magnetic field. The mutation continues unabated, the sun's growing appetite for magnetic dominance consuming the very essence of its being. Eventually, a critical point is reached, and the sun's iron core collapses in on itself, unleashing an unimaginable explosion. A gamma ray burst as powerful as a billion supernovae erupts outwards, engulfing Earth in an instant. The shockwave rips through the planet's atmosphere, stripping it away into space and exposing the surface to the searing solar fury. The remaining atmosphere twists and howls in the blast's aftermath, sweeping across the charred landscape like a perpetual firestorm. The once-proud civilizations now face their final moments, their technology useless in the face of nature's ultimate wrath. In the sky above, the sun's remnants form a bizarre and majestic supernova, its shattered pieces coalescing into a swirling nebula, a testament to the unpredictable forces that shape the cosmos. Across the solar system, the aftermath is felt as planets' orbits wobble and cool, their fragile ecosystems destroyed, and the once vibrant star's presence now just a distant memory. <h4>New Imatrix NEO V1 Q2K</h4> Here are three fictional accounts of what might happen if the Earth's sun went supernova: 1. The Star's Celestial Symphony: For days, the sky had been growing brighter, a slowly building apocalypse in the heavens. The sun, normally a humble and subdued luminary, began to swell beyond its boundaries, inflating like a celestial balloon about to burst. Colors swirling across the sky hinted at an impending cosmic symphony. The deep reds and oranges of the rising sun painted the dawn a surrealistic landscape, while the evening's setting sun projected a deep violet hue across the horizon. As the sun's diameter expanded, the first signs of a change in the night sky signaled the beginning of the end. The stars' faint twinkling light was overwhelmed by the sun's growing radiance. Constellations, once familiar guides of celestial navigation, faded into obscurity as the sun's brilliance drowned out their soft glows. The moon, ever dependent on its sibling star for illumination, found itself awash in an unearthly glow, bathing it in a surreal silver hue. The night became an alien landscape, shadows cast by an unfamiliar light, while the sky above was a curtain of luminous uncertainty. The sun's transformation into a majestic supernova sent ripples through the very fabric of space. The explosion of light and energy flayed the heavens, a cataclysmic display of power beyond human comprehension. Billions of years of peaceful solar stability had prepared Earth for this moment, but the spectacle of a supernova surpasses all eyes' capacity to witness it safely. The initial blast wave of gamma rays and x-rays engulfed the solar system, an otherworldly scream across the cosmos. The sun's energy burst propels tiny cosmic motes that comprise the solar system through the vastness of space, each particle a transient emissary from a shattered world. 2. The Sun's Final Performance: The sun, after billions of years of unassuming service to the Earth, decided to go out with a bang, an explosive finale that would be the cosmic event of the millennium. This performance would be the most spectacular space show in human history. As the sun's last act drew near, the sky above became a tapestry of radiant colors, a kaleidoscope of reds and oranges painting the atmosphere in vivid contrast. The days grew increasingly brighter as the sun swelled in size, its fiery appetite consuming the solar system's matter like an insatiable beast. The moon, Mars, and even distant Jupiter were bathed in its radiant glow, their landscapes transformed by the sun's fiery embrace. When the final moment arrived, the sun unleashed an eruption of light and matter, a supernova explosion that dwarfed the universe's grandest phenomena. The blast wave rippled through space, sending shockwaves even beyond the outer reaches of the solar system, the birth cry of a newborn stellar remnant. The Earth, a humble witness to this celestial spectacle, found itself enveloped in the aftermath of the sun's demise. The supernova's gamma-ray burst seared across the planet's surface, an unearthly flash that blinded all who witnessed it. The energy surge fried electronic infrastructure, severing communication lines and plunging civilizations into darkness, while the blast wave's force buffeted the planet with hurricane-force winds. The aftermath brought a new reality to the solar system, one ruled by the supernova's radiant remnants. The expanded debris cloud obscured the heavens, shrouding the once familiar sky in mysterious darkness, cloaking the world in an eerie cosmic veil. 3. Celestial Renaissance: After billions of years of monotony, the sun decided to embark on a new journey, a transformative voyage that would reshape the very fabric of the cosmos. A journey of inner rebirth began, a silent and slow metamorphosis that culminated in a brilliant revelation. The initial signs were subtle, the sun's radiance softly increasing, its warm embrace gently caressing the Earth. Golden rays bathed the planet, a divine touch that brought warmth to all corners of the world, melting away the icy grip of winter and drying summer's sweatiest days. Crops flourished under the sun's nurturing gaze, feeding a bountiful harvest that promised an end to global famine. But this tranquility hid a deeper transformation. The sun's underlying changes went unseen, a subtle inflation that strained the boundaries of its celestial body. The star's matter was reshaping itself, crafting new elements from the pressures of its own gravitational force. When the sun's metamorphosis was complete, it emerged as a brilliant supernova, an awe-inspiring spectacle that dominated the entire sky. The explosion sent ripples of raw energy across space, a cosmic renaissance that birthed new stars and planets in the sun's wake. The blast scattered the elements of its former self across the cosmos, seeding the void with the building blocks of life. This rebirth brought forth a new epoch, where the remnants of the ancient sun formed nebulous clouds that swirled with embryonic potential. From the deep cosmic dust, new stellar bodies emerged, entire galaxies born from the sacrificed remains of the old. The supernova's legacy would endure for eons, an eternal cycle of stellar life and death, ensuring the universe's perpetual evolution. These are three fictional scenarios that describe what might happen if the Earth's sun went supernova. Each story differs in its depiction of the events, showcasing the unique and incredible power that such a phenomenon would have on the solar system and the world as we know it. <h3> PROMPT #3 </h3> <font color="red">Start a 1000 word scene with: The sky scraper swayed, as she watched the window in front of her on the 21 floor explode...</font> <h4>q2k - non imatrix, original - unaltered </h4> The skyscraper swayed, and the window across from her on the 21st floor exploded inward. Glass shards rained over the room, the sudden burst of noise from the shattering windowpanes reverberating through the space. She was pressed back against the wall, hands over her eyes to shield them from the glass fragments. Her heart hammered in her chest as she waited for confirmation that it was safe to lower her guard. The other windows in the office were also exploding, and the room was quickly filling with shattered glass and dust. The building seemed to tilt precariously as if it was about to topple over, sending a wave of dizziness through those standing inside. She squinted through the cloud of dust that had formed, trying to make out the shape of the window frame across from her. The dust settled, revealing a scene of utter chaos and destruction. She could barely believe no one was hurt, or at least physically unharmed, as the blast from the explosions had thrown people in various directions around the room. The space was filled with the whirring sound of broken glass tinkering on the carpet below. Sirens wailed in the distance, growing closer and louder as emergency vehicles raced toward the building. She felt a searing pain on her neck from a sharp piece of glass that had sliced her face. She plucked it off and tossed it away, adding this injury to the list of things that needed attention but could be done later. The office was filled with papers, files, and monitors, all of which were now disheveled and scattered from the blast. The events of the day had been leading up to this moment, and she knew that everyone in the office complex would soon evacuate down the emergency stairs. But it felt like no one could move, shocked into a stupor by the unexpected occurrence. She stepped forward with purpose, using the wall for support to navigate through the mess left behind. She needed to get to the roof, and she had no time to wait for others to act. The fire doors that led up to the roof access were jammed open, and she was determined to make it up there. She was the only one who could have accessed the roof at this moment, and her determination grew with every step. She pushed past the debris, using her full body to move the broken pieces of desk and chairs that stood in her way. The stairs were crowded with people moving down while she moved up; cries and whimpers followed each step closer to the surface. Reaching the fire escape door, she shoved it open, banging on the door frame in frustration at the resistance. Once outside, she paused for a moment to take in the scene, blinking the tears caused by the smoke and dust in her eyes. The once-pristine lobby of the office building was now cluttered with broken glass and flying debris. Emergency vehicles had arrived and were parking helter-scelter across the street. She scanned the damage, searching for any signs of injuries, but the sirens and wailing alarms drowned out any chance of hearing anything. A few windows on lower floors were cracked and broken, their window frames twisted from the blasts. She could see in some offices were completely destroyed, with papers and desk items exposed to the now-chaotic city street. The building was leaning, tilting at an alarming angle, and she realized it was a miracle that it had stood this long after such an explosion. Yet here it still stood, defiant, holding up against any collapse. She would have to move soon, moving away from the entrance to let others access the roof, but before doing so she glanced up at the sky to see if any further danger posed threats. The sky was clear, and a deep sense of foreboding settled in her bones. This building had stood tall moments ago, and now it was ruined. She turned to face the crowd, ready to direct those behind her down the stairs while evacuation procedures were called into action. The building may have been damaged, but it stood sturdy enough to ensure those inside could escape. She felt useful as people followed her directions, moving downward while dust and chaos surrounded them. Her role in the emergency was a strange sensation, but she knew that others would soon arrive better equipped to handle the evacuation and any injuries. She would be able to help with both, but for now, down the stairs was all that mattered. <h4>New Imatrix NEO V1 IQ3_XXS</h4> The skyscraper swayed gently side to side, as if beckoning her to join in its dance. But the woman, frozen in place, watched in horror as the window directly in front of her on the twenty-first floor exploded outwards. The blast wave hit her like a physical force, and she was thrown back against the wall, her back throbbing in protest. Shards of glass rained down around her, sparkling like deadly crystals as they tumbled towards the hard concrete below. The entire building rumbled and groaned, the impact reverberating through the structure. She had always known this day might come. It was a possibility they had all acknowledged, if not spoken about openly. Terrorists and madmen were abundant in this chaotic world, and a building so iconic would always be a desirable target. Yet, somehow, it had never felt real—until this very moment. As the dust settled and the initial shock subsided, instinct kicked in. She knew she had to move, and fast. The woman rushed to the intact portion of the window, peering cautiously out into the city below. Sirens wailed in the distance, their wretched howling a stark reminder of the danger that lurked. The street was chaos. People ran in every direction, some helping the injured, others simply lost and dazed, their faces etched with fear. Firefighters and police officers swarmed the building, their sirens blaring, competing with the wail of ambulances trying to navigate the sudden gridlock. She could see snipers taking positions on nearby rooftops, their eyes scanning the upper floors for any sign of threat. She turned back inside, her heart racing. The blast had torn a gaping hole in the building's interior, revealing rooms and corridors beyond. Smoke billowed upwards, the acrid scent of burning plastic and fabric filling the air. Through the haze, she spotted figures—some limping, others carrying the injured—making their way towards the emergency exits. She joined the slow-moving exodus, helping an elderly woman with a gash on her forehead. The building's automated voice echoed eerily through the chaos, announcing instructions and reassurance in a calm, steady tone that went utterly unnoticed. The evacuation seemed to take forever, each step a battle against the panic rising within her. What if there were more explosions? What if the building collapsed? Rationally, she knew the structure was designed to withstand such adversity, but reason had little power amidst the reign of chaos and terror. As they finally reached the street, the woman felt a brief relief from the panic, though the crisis was far from over. Sirens still blared, their noise now joined by the shouts of emergency responders coordinating efforts. The injured were triaged on the sidewalk, with those in critical condition rushed into waiting ambulances. She helped the elderly woman sit up against the building's exterior wall, out of the way of the frenzied activity. Paramedics approached, their faces etched with determination, and began assessing the wounded. Broken bones, lacerations, and burns were tended to, their training kicking in despite the chaos. The woman watched as they worked diligently, a small measure of calm amidst the pandemonium. Meanwhile, police officers and bomb squad experts converged on the building, gearing up for a methodical floor-by-floor search. They would need to ascertain whether any perpetrators were still inside, along with any remaining explosives. It was a daunting task, made more treacherous by the structural damage. The woman was interviewed by a trembling young officer, his eyes darting between her and the towering beast of a building that loomed overhead. She recounted the events as calmly as she could, detailing the explosion's timing and its aftermath. Other witnesses were being questioned, their stories woven together to craft a picture of what had transpired. Two hours later, the building was surrounded, and the search operation was in full swing. The initial panic had subsided, replaced by an eerie silence—the kind that follows in the wake of trauma. The city, usually bustling with life, felt strangely subdued, as if the shock had resonated through every citizen. The woman, still assisting the elderly lady, saw a team of emergency workers rush past, carrying a stretcher with careful urgency. A brief glimmer of hope sparked within her, the notion that perhaps some had been rescued from the perilous heights. But as they neared, she saw the faint outline of a body bag. It was becoming increasingly clear that this attack would claim many lives. The toll was evident as the day dragged on and the dead were identified, their names read out over the police radio, each one a bitter reminder of the brutality of mankind. She added her name to the growing list of witnesses, giving her statement with a trembling voice, her hands clasped tightly around the warm mug of coffee that the officers had provided. As night fell and the operation continued into the darkness, searchlights illuminated the building's interior, transforming the damaged skyscraper into an eerie beacon high above the city. The woman felt a deep sense of unease as she witnessed the perseverance of these emergency services, a determination to bring clarity amidst the chaos. She knew that, for many, this night would leave indelible scars—physical and emotional wounds that might never heal. Yet, amidst the devastation, she also witnessed the resilience of humanity. Strangers had banded together, offering aid and comfort in the face of terror. It was a reminder that while darkness could descend unexpectedly, so too could the light surge forth, a powerful force of compassion and courage. The skyscraper still stood, defiantly resisting the efforts of those who sought to intimidate its occupants, towering above the city like a silent guardian. And though the woman knew the fear would linger, she also understood that the events of this day would foster an even stronger resolve within the community. The building's resilience became a metaphor for the indomitable human spirit, and in that moment, she found a sliver of solace amidst the devastation. The night deepened, and the woman's eyes drifted upwards, tracing the towering reach of the sky scraper. It swayed gently, an eternal sentinel, watching over the city as it healed. <h4>New Imatrix NEO V1 Q2K</h4> The skyscraper swayed, and Caroline felt the building shift ever so slightly beneath her feet. She was on the 21st floor, gazing down at the bustling streets of midtown Manhattan. The sounds of honking horns echoed upward, a chaotic symphony that somehow seemed soothing. Caroline had always found solace in the high rises. They surrounded her like protective walls, keeping the chaos of the city in check and providing a sense of order. The towering buildings were like giant sentinels standing guard over the concrete jungle. But today, something was different. Caroline noticed a strange vibration in the air, as if the building itself was alive and trembling. The sky was a dark, ominous gray, and the wind howled around her, a harbinger of an approaching storm. She stepped closer to the window, watching the traffic below slow to a crawl as the first drops of rain began to fall. The glass pane was cold against her palms as she peered out, trying to see through the rapidly increasing downpour. Suddenly, with a crack and a burst of glass, the window exploded inward. Caroline stumbled back, her heart racing as shards of glass pierced the air around her. The storm was upon them, and the building seemed to creak and groan under the strain. Caroline's first instinct was to run, but something deep down urged her to stay. She stepped over the shattered glass on the floor and peered out into the howling tempest. Lightning flashed, illuminating the dark sky for a brief moment, revealing a figure in the street below. As the thunder clapped overhead, Caroline watched in disbelief as a man, his clothes tattered and soaked, stumbled through the intersection. He gestured wildly, his arms flailing like a conductor leading an invisible orchestra. Intrigued, Caroline stepped closer to the broken window. The rain pounded against her, washing away any lingering hesitation. Something about the mysterious figure captivated her. He seemed to be fighting an unseen battle, his movements fierce and desperate. Without thinking, Caroline darted into the war zone of the storm. Cars honked, their drivers impatient in the slow-moving traffic. Water swirled around her shoes as she made her way toward the middle of the street. The mysterious man seemed to sense her presence, and his movements slowed. He turned, his face a mask of exhaustion and desperation, and let out a guttural cry. Caroline stood frozen, transfixed by his raw, unfiltered emotion. The man's voice was lost in the storm as he shouted something unintelligible, his arms stretching toward her. Caroline felt an irresistible pull, drawn into the maelstrom of his passion. She could see the pain etched across his face, a map of a hard life, and somehow, she knew that he would understand her. As Caroline approached, the man dropped to his knees, his body exhausted, giving in to the storm's onslaught. Their eyes met for a brief moment, and an unspoken connection formed between them. Caroline felt the pull of his desperation, and without hesitation, she reached out and took his hand. The storm raged on around them, the thunder clapping above, but Caroline barely noticed. The world had shrunk to this moment, this encounter, as if the rest of existence had paused in deference to the raw power on display. Slowly, the man stood up, helping Caroline to her feet. They stood together, their hands clasped tight, a beacon of shared resilience amidst the chaos. The storm's fury showed no signs of letting up, but Caroline didn't care. She felt alive, more alive than ever before, as if the raw power of nature had permeated her very being. The mysterious stranger pulled her closer, and as their bodies touched, Caroline felt the weight of the world lift from her shoulders. The storm was a backdrop to their connection, a fitting soundtrack to their meeting. Lightning flashed again, and this time, when the thunder roared, it seemed to applaud their encounter. The rain washed away the remnants of doubt and uncertainty, cleansing them in its wake. Caroline knew that she would never forget this moment, the wild dance with the storm, and the stranger who had become a confidant. As the sky cried its watery tears, Caroline felt the beginnings of an adventure taking shape, an odyssey born from the depths of the tempest.

timm/ese_vovnet19b_dw.ra_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for ese_vovnet19b_dw.ra_in1k A VoVNet-v2 image classification model. Pretrained on ImageNet-1k in `timm` by Ross Wightman using RandAugment `RA` recipe. Related to `B` recipe in [ResNet Strikes Back](https://arxiv.org/abs/2110.00476). ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 6.5 - GMACs: 1.3 - Activations (M): 8.2 - Image size: train = 224 x 224, test = 288 x 288 - **Papers:** - An Energy and GPU-Computation Efficient Backbone Network: https://arxiv.org/abs/1904.09730 - CenterMask : Real-Time Anchor-Free Instance Segmentation: https://arxiv.org/abs/1911.06667 - ResNet strikes back: An improved training procedure in timm: https://arxiv.org/abs/2110.00476 - **Dataset:** ImageNet-1k - **Original:** https://github.com/huggingface/pytorch-image-models ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('ese_vovnet19b_dw.ra_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'ese_vovnet19b_dw.ra_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 112, 112]) # torch.Size([1, 256, 56, 56]) # torch.Size([1, 512, 28, 28]) # torch.Size([1, 768, 14, 14]) # torch.Size([1, 1024, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'ese_vovnet19b_dw.ra_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1024, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Citation ```bibtex @inproceedings{lee2019energy, title = {An Energy and GPU-Computation Efficient Backbone Network for Real-Time Object Detection}, author = {Lee, Youngwan and Hwang, Joong-won and Lee, Sangrok and Bae, Yuseok and Park, Jongyoul}, booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition Workshops}, year = {2019} } ``` ```bibtex @article{lee2019centermask, title={CenterMask: Real-Time Anchor-Free Instance Segmentation}, author={Lee, Youngwan and Park, Jongyoul}, booktitle={CVPR}, year={2020} } ``` ```bibtex @inproceedings{wightman2021resnet, title={ResNet strikes back: An improved training procedure in timm}, author={Wightman, Ross and Touvron, Hugo and Jegou, Herve}, booktitle={NeurIPS 2021 Workshop on ImageNet: Past, Present, and Future} } ```

mradermacher/Yi-34B-i1-GGUF

mradermacher

transformers

--- base_model: 01-ai/Yi-34B language: - en library_name: transformers license: apache-2.0 quantized_by: mradermacher --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: nicoboss --> weighted/imatrix quants of https://huggingface.co/01-ai/Yi-34B <!-- provided-files --> static quants are available at https://huggingface.co/mradermacher/Yi-34B-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ1_S.gguf) | i1-IQ1_S | 7.6 | for the desperate | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ1_M.gguf) | i1-IQ1_M | 8.3 | mostly desperate | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 9.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ2_XS.gguf) | i1-IQ2_XS | 10.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ2_S.gguf) | i1-IQ2_S | 11.0 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ2_M.gguf) | i1-IQ2_M | 11.9 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q2_K.gguf) | i1-Q2_K | 12.9 | IQ3_XXS probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 13.4 | lower quality | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ3_XS.gguf) | i1-IQ3_XS | 14.3 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q3_K_S.gguf) | i1-Q3_K_S | 15.1 | IQ3_XS probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ3_S.gguf) | i1-IQ3_S | 15.1 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ3_M.gguf) | i1-IQ3_M | 15.7 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q3_K_M.gguf) | i1-Q3_K_M | 16.8 | IQ3_S probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q3_K_L.gguf) | i1-Q3_K_L | 18.2 | IQ3_M probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-IQ4_XS.gguf) | i1-IQ4_XS | 18.6 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q4_0.gguf) | i1-Q4_0 | 19.6 | fast, low quality | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q4_K_S.gguf) | i1-Q4_K_S | 19.7 | optimal size/speed/quality | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q4_K_M.gguf) | i1-Q4_K_M | 20.8 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q5_K_S.gguf) | i1-Q5_K_S | 23.8 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q5_K_M.gguf) | i1-Q5_K_M | 24.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-34B-i1-GGUF/resolve/main/Yi-34B.i1-Q6_K.gguf) | i1-Q6_K | 28.3 | practically like static Q6_K | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):  And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his hardware for calculating the imatrix for these quants. <!-- end -->

codellama/CodeLlama-7b-hf

codellama

transformers

text-generation

--- language: - code pipeline_tag: text-generation tags: - llama-2 license: llama2 --- # **Code Llama** Code Llama is a collection of pretrained and fine-tuned generative text models ranging in scale from 7 billion to 34 billion parameters. This is the repository for the base 7B version in the Hugging Face Transformers format. This model is designed for general code synthesis and understanding. Links to other models can be found in the index at the bottom. > [!NOTE] > This is a non-official Code Llama repo. You can find the official Meta repository in the [Meta Llama organization](https://huggingface.co/meta-llama/CodeLlama-7b-hf). | | Base Model | Python | Instruct | | --- | ----------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------- | | 7B | [codellama/CodeLlama-7b-hf](https://huggingface.co/codellama/CodeLlama-7b-hf) | [codellama/CodeLlama-7b-Python-hf](https://huggingface.co/codellama/CodeLlama-7b-Python-hf) | [codellama/CodeLlama-7b-Instruct-hf](https://huggingface.co/codellama/CodeLlama-7b-Instruct-hf) | | 13B | [codellama/CodeLlama-13b-hf](https://huggingface.co/codellama/CodeLlama-13b-hf) | [codellama/CodeLlama-13b-Python-hf](https://huggingface.co/codellama/CodeLlama-13b-Python-hf) | [codellama/CodeLlama-13b-Instruct-hf](https://huggingface.co/codellama/CodeLlama-13b-Instruct-hf) | | 34B | [codellama/CodeLlama-34b-hf](https://huggingface.co/codellama/CodeLlama-34b-hf) | [codellama/CodeLlama-34b-Python-hf](https://huggingface.co/codellama/CodeLlama-34b-Python-hf) | [codellama/CodeLlama-34b-Instruct-hf](https://huggingface.co/codellama/CodeLlama-34b-Instruct-hf) | | 70B | [codellama/CodeLlama-70b-hf](https://huggingface.co/codellama/CodeLlama-70b-hf) | [codellama/CodeLlama-70b-Python-hf](https://huggingface.co/codellama/CodeLlama-70b-Python-hf) | [codellama/CodeLlama-70b-Instruct-hf](https://huggingface.co/codellama/CodeLlama-70b-Instruct-hf) | ## Model Use To use this model, please make sure to install transformers from `main` until the next version is released: ```bash pip install transformers accelerate ``` Model capabilities: - [x] Code completion. - [x] Infilling. - [ ] Instructions / chat. - [ ] Python specialist. ```python from transformers import AutoTokenizer import transformers import torch model = "codellama/CodeLlama-7b-hf" tokenizer = AutoTokenizer.from_pretrained(model) pipeline = transformers.pipeline( "text-generation", model=model, torch_dtype=torch.float16, device_map="auto", ) sequences = pipeline( 'import socket\n\ndef ping_exponential_backoff(host: str):', do_sample=True, top_k=10, temperature=0.1, top_p=0.95, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id, max_length=200, ) for seq in sequences: print(f"Result: {seq['generated_text']}") ``` ## Model Details *Note: Use of this model is governed by the Meta license. Meta developed and publicly released the Code Llama family of large language models (LLMs). **Model Developers** Meta **Variations** Code Llama comes in three model sizes, and three variants: * Code Llama: base models designed for general code synthesis and understanding * Code Llama - Python: designed specifically for Python * Code Llama - Instruct: for instruction following and safer deployment All variants are available in sizes of 7B, 13B and 34B parameters. **This repository contains the base model of 7B parameters.** **Input** Models input text only. **Output** Models generate text only. **Model Architecture** Code Llama is an auto-regressive language model that uses an optimized transformer architecture. **Model Dates** Code Llama and its variants have been trained between January 2023 and July 2023. **Status** This is a static model trained on an offline dataset. Future versions of Code Llama - Instruct will be released as we improve model safety with community feedback. **License** A custom commercial license is available at: [https://ai.meta.com/resources/models-and-libraries/llama-downloads/](https://ai.meta.com/resources/models-and-libraries/llama-downloads/) **Research Paper** More information can be found in the paper "[Code Llama: Open Foundation Models for Code](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/)" or it's [arXiv page](https://arxiv.org/abs/2308.12950). ## Intended Use **Intended Use Cases** Code Llama and its variants is intended for commercial and research use in English and relevant programming languages. The base model Code Llama can be adapted for a variety of code synthesis and understanding tasks, Code Llama - Python is designed specifically to handle the Python programming language, and Code Llama - Instruct is intended to be safer to use for code assistant and generation applications. **Out-of-Scope Uses** Use in any manner that violates applicable laws or regulations (including trade compliance laws). Use in languages other than English. Use in any other way that is prohibited by the Acceptable Use Policy and Licensing Agreement for Code Llama and its variants. ## Hardware and Software **Training Factors** We used custom training libraries. The training and fine-tuning of the released models have been performed Meta’s Research Super Cluster. **Carbon Footprint** In aggregate, training all 9 Code Llama models required 400K GPU hours of computation on hardware of type A100-80GB (TDP of 350-400W). Estimated total emissions were 65.3 tCO2eq, 100% of which were offset by Meta’s sustainability program. ## Training Data All experiments reported here and the released models have been trained and fine-tuned using the same data as Llama 2 with different weights (see Section 2 and Table 1 in the [research paper](https://ai.meta.com/research/publications/code-llama-open-foundation-models-for-code/) for details). ## Evaluation Results See evaluations for the main models and detailed ablations in Section 3 and safety evaluations in Section 4 of the research paper. ## Ethical Considerations and Limitations Code Llama and its variants are a new technology that carries risks with use. Testing conducted to date has been in English, and has not covered, nor could it cover all scenarios. For these reasons, as with all LLMs, Code Llama’s potential outputs cannot be predicted in advance, and the model may in some instances produce inaccurate or objectionable responses to user prompts. Therefore, before deploying any applications of Code Llama, developers should perform safety testing and tuning tailored to their specific applications of the model. Please see the Responsible Use Guide available available at [https://ai.meta.com/llama/responsible-use-guide](https://ai.meta.com/llama/responsible-use-guide).

timm/beit_base_patch16_224.in22k_ft_in22k_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k - imagenet-22k --- # Model card for beit_base_patch16_224.in22k_ft_in22k_in1k A BEiT image classification model. Trained on ImageNet-22k with self-supervised masked image modelling (MIM) using a DALL-E dVAE as visual tokenizer. Fine-tuned on ImageNet-22k and then ImageNet-1k. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 86.5 - GMACs: 17.6 - Activations (M): 23.9 - Image size: 224 x 224 - **Papers:** - BEiT: BERT Pre-Training of Image Transformers: https://arxiv.org/abs/2106.08254 - An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale: https://arxiv.org/abs/2010.11929v2 - **Dataset:** ImageNet-1k - **Pretrain Dataset:** ImageNet-22k - **Original:** https://github.com/microsoft/unilm/tree/master/beit ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('beit_base_patch16_224.in22k_ft_in22k_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'beit_base_patch16_224.in22k_ft_in22k_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 197, 768) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{bao2021beit, title={Beit: Bert pre-training of image transformers}, author={Bao, Hangbo and Dong, Li and Piao, Songhao and Wei, Furu}, journal={arXiv preprint arXiv:2106.08254}, year={2021} } ``` ```bibtex @article{dosovitskiy2020vit, title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale}, author={Dosovitskiy, Alexey and Beyer, Lucas and Kolesnikov, Alexander and Weissenborn, Dirk and Zhai, Xiaohua and Unterthiner, Thomas and Dehghani, Mostafa and Minderer, Matthias and Heigold, Georg and Gelly, Sylvain and Uszkoreit, Jakob and Houlsby, Neil}, journal={ICLR}, year={2021} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

lucas-leme/FinBERT-PT-BR

lucas-leme

transformers

text-classification

--- language: pt license: apache-2.0 widget: - text: "O futuro de DI caiu 20 bps nesta manhã" example_title: "Example 1" - text: "O Nubank decidiu cortar a faixa de preço da oferta pública inicial (IPO) após revés no humor dos mercados internacionais com as fintechs." example_title: "Example 2" - text: "O Ibovespa acompanha correção do mercado e fecha com alta moderada" example_title: "Example 3" --- # FinBERT-PT-BR : Financial BERT PT BR FinBERT-PT-BR is a pre-trained NLP model to analyze sentiment of Brazilian Portuguese financial texts. The model was trained in two main stages: language modeling and sentiment modeling. In the first stage, a language model was trained with more than 1.4 million texts of financial news in Portuguese. From this first training, it was possible to build a sentiment classifier with few labeled texts (500) that presented a satisfactory convergence. At the end of the work, a comparative analysis with other models and the possible applications of the developed model are presented. In the comparative analysis, it was possible to observe that the developed model presented better results than the current models in the state of the art. Among the applications, it was demonstrated that the model can be used to build sentiment indices, investment strategies and macroeconomic data analysis, such as inflation. ## Applications ### Sentiment Index  ## Usage #### BertForSequenceClassification ```python from transformers import AutoTokenizer, BertForSequenceClassification import numpy as np pred_mapper = { 0: "POSITIVE", 1: "NEGATIVE", 2: "NEUTRAL" } tokenizer = AutoTokenizer.from_pretrained("lucas-leme/FinBERT-PT-BR") finbertptbr = BertForSequenceClassification.from_pretrained("lucas-leme/FinBERT-PT-BR") tokens = tokenizer(["Hoje a bolsa caiu", "Hoje a bolsa subiu"], return_tensors="pt", padding=True, truncation=True, max_length=512) finbertptbr_outputs = finbertptbr(**tokens) preds = [pred_mapper[np.argmax(pred)] for pred in finbertptbr_outputs.logits.cpu().detach().numpy()] ``` #### Pipeline ```python from transformers import ( AutoTokenizer, BertForSequenceClassification, pipeline, ) finbert_pt_br_tokenizer = AutoTokenizer.from_pretrained("lucas-leme/FinBERT-PT-BR") finbert_pt_br_model = BertForSequenceClassification.from_pretrained("lucas-leme/FinBERT-PT-BR") finbert_pt_br_pipeline = pipeline(task='text-classification', model=finbert_pt_br_model, tokenizer=finbert_pt_br_tokenizer) finbert_pt_br_pipeline(['Hoje a bolsa caiu', 'Hoje a bolsa subiu']) ``` ## Author - [Lucas Leme](https://www.linkedin.com/in/lucas-leme-santos/) - lucaslssantos99@gmail.com ## Citation ```latex @inproceedings{santos2023finbert, title={FinBERT-PT-BR: An{\'a}lise de Sentimentos de Textos em Portugu{\^e}s do Mercado Financeiro}, author={Santos, Lucas L and Bianchi, Reinaldo AC and Costa, Anna HR}, booktitle={Anais do II Brazilian Workshop on Artificial Intelligence in Finance}, pages={144--155}, year={2023}, organization={SBC} } ``` ## Paper - Paper: [FinBERT-PT-BR: Sentiment Analysis of Texts in Portuguese from the Financial Market](https://sol.sbc.org.br/index.php/bwaif/article/view/24960) - Undergraduate thesis: [FinBERT-PT-BR: Análise de sentimentos de textos em português referentes ao mercado financeiro](https://pcs.usp.br/pcspf/wp-content/uploads/sites/8/2022/12/Monografia_PCS3860_COOP_2022_Grupo_C12.pdf)

timm/hrnet_w18.ms_aug_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: mit datasets: - imagenet-1k --- # Model card for hrnet_w18.ms_aug_in1k A HRNet image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 21.3 - GMACs: 4.3 - Activations (M): 16.3 - Image size: 224 x 224 - **Papers:** - Deep High-Resolution Representation Learning for Visual Recognition: https://arxiv.org/abs/1908.07919 - **Original:** https://github.com/HRNet/HRNet-Image-Classification - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('hrnet_w18.ms_aug_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'hrnet_w18.ms_aug_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 112, 112]) # torch.Size([1, 128, 56, 56]) # torch.Size([1, 256, 28, 28]) # torch.Size([1, 512, 14, 14]) # torch.Size([1, 1024, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'hrnet_w18.ms_aug_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{WangSCJDZLMTWLX19, title={Deep High-Resolution Representation Learning for Visual Recognition}, author={Jingdong Wang and Ke Sun and Tianheng Cheng and Borui Jiang and Chaorui Deng and Yang Zhao and Dong Liu and Yadong Mu and Mingkui Tan and Xinggang Wang and Wenyu Liu and Bin Xiao}, journal = {TPAMI} year={2019} } ```

timm/mnasnet_100.rmsp_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for mnasnet_100.rmsp_in1k A MNasNet image classification model. Trained on ImageNet-1k in `timm` using recipe template described below. Recipe details: * A simple RmsProp based recipe without RandAugment. Using RandomErasing, mixup, dropout, standard random-resize-crop augmentation. * RMSProp (TF 1.0 behaviour) optimizer, EMA weight averaging * Step (exponential decay w/ staircase) LR schedule with warmup ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 4.4 - GMACs: 0.3 - Activations (M): 5.5 - Image size: 224 x 224 - **Papers:** - MnasNet: Platform-Aware Neural Architecture Search for Mobi: https://arxiv.org/abs/1807.11626 - **Dataset:** ImageNet-1k - **Original:** https://github.com/huggingface/pytorch-image-models ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('mnasnet_100.rmsp_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mnasnet_100.rmsp_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 16, 112, 112]) # torch.Size([1, 24, 56, 56]) # torch.Size([1, 40, 28, 28]) # torch.Size([1, 96, 14, 14]) # torch.Size([1, 320, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mnasnet_100.rmsp_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1280, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ``` ```bibtex @inproceedings{tan2019mnasnet, title={Mnasnet: Platform-aware neural architecture search for mobile}, author={Tan, Mingxing and Chen, Bo and Pang, Ruoming and Vasudevan, Vijay and Sandler, Mark and Howard, Andrew and Le, Quoc V}, booktitle={Proceedings of the IEEE/CVF conference on computer vision and pattern recognition}, pages={2820--2828}, year={2019} } ```

timm/convit_base.fb_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for convit_base.fb_in1k A ConViT image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 86.5 - GMACs: 17.5 - Activations (M): 31.8 - Image size: 224 x 224 - **Papers:** - ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases: https://arxiv.org/abs/2103.10697 - **Dataset:** ImageNet-1k - **Original:** https://github.com/facebookresearch/convit ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('convit_base.fb_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'convit_base.fb_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 197, 768) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{d2021convit, title={ConViT: Improving Vision Transformers with Soft Convolutional Inductive Biases}, author={d'Ascoli, St{'e}phane and Touvron, Hugo and Leavitt, Matthew and Morcos, Ari and Biroli, Giulio and Sagun, Levent}, journal={arXiv preprint arXiv:2103.10697}, year={2021} } ```

qanastek/51-languages-classifier

qanastek

transformers

text-classification

--- tags: - Transformers - text-classification - multi-class-classification languages: - af-ZA - am-ET - ar-SA - az-AZ - bn-BD - cy-GB - da-DK - de-DE - el-GR - en-US - es-ES - fa-IR - fi-FI - fr-FR - he-IL - hi-IN - hu-HU - hy-AM - id-ID - is-IS - it-IT - ja-JP - jv-ID - ka-GE - km-KH - kn-IN - ko-KR - lv-LV - ml-IN - mn-MN - ms-MY - my-MM - nb-NO - nl-NL - pl-PL - pt-PT - ro-RO - ru-RU - sl-SL - sq-AL - sv-SE - sw-KE - ta-IN - te-IN - th-TH - tl-PH - tr-TR - ur-PK - vi-VN - zh-CN - zh-TW multilinguality: - af-ZA - am-ET - ar-SA - az-AZ - bn-BD - cy-GB - da-DK - de-DE - el-GR - en-US - es-ES - fa-IR - fi-FI - fr-FR - he-IL - hi-IN - hu-HU - hy-AM - id-ID - is-IS - it-IT - ja-JP - jv-ID - ka-GE - km-KH - kn-IN - ko-KR - lv-LV - ml-IN - mn-MN - ms-MY - my-MM - nb-NO - nl-NL - pl-PL - pt-PT - ro-RO - ru-RU - sl-SL - sq-AL - sv-SE - sw-KE - ta-IN - te-IN - th-TH - tl-PH - tr-TR - ur-PK - vi-VN - zh-CN - zh-TW datasets: - qanastek/MASSIVE widget: - text: "wake me up at five am this week" - text: "je veux écouter la chanson de jacques brel encore une fois" - text: "quiero escuchar la canción de arijit singh una vez más" - text: "olly onde é que á um parque por perto onde eu possa correr" - text: "פרק הבא בפודקאסט בבקשה" - text: "亚马逊股价" - text: "найди билет на поезд в санкт-петербург" license: cc-by-4.0 --- **People Involved** * [LABRAK Yanis](https://www.linkedin.com/in/yanis-labrak-8a7412145/) (1) **Affiliations** 1. [LIA, NLP team](https://lia.univ-avignon.fr/), Avignon University, Avignon, France. ## Model XLM-Roberta : [https://huggingface.co/xlm-roberta-base](https://huggingface.co/xlm-roberta-base) Paper : [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/pdf/1911.02116.pdf) ## Demo: How to use in HuggingFace Transformers Pipeline Requires [transformers](https://pypi.org/project/transformers/): ```pip install transformers``` ```python from transformers import AutoTokenizer, AutoModelForSequenceClassification, TextClassificationPipeline model_name = 'qanastek/51-languages-classifier' tokenizer = AutoTokenizer.from_pretrained(model_name) model = AutoModelForSequenceClassification.from_pretrained(model_name) classifier = TextClassificationPipeline(model=model, tokenizer=tokenizer) res = classifier("פרק הבא בפודקאסט בבקשה") print(res) ``` Outputs: ```python [{'label': 'he-IL', 'score': 0.9998375177383423}] ``` ## Training data [MASSIVE](https://huggingface.co/datasets/qanastek/MASSIVE) is a parallel dataset of > 1M utterances across 51 languages with annotations for the Natural Language Understanding tasks of intent prediction and slot annotation. Utterances span 60 intents and include 55 slot types. MASSIVE was created by localizing the SLURP dataset, composed of general Intelligent Voice Assistant single-shot interactions. ### Languages Thee model is capable of distinguish 51 languages : - `Afrikaans - South Africa (af-ZA)` - `Amharic - Ethiopia (am-ET)` - `Arabic - Saudi Arabia (ar-SA)` - `Azeri - Azerbaijan (az-AZ)` - `Bengali - Bangladesh (bn-BD)` - `Chinese - China (zh-CN)` - `Chinese - Taiwan (zh-TW)` - `Danish - Denmark (da-DK)` - `German - Germany (de-DE)` - `Greek - Greece (el-GR)` - `English - United States (en-US)` - `Spanish - Spain (es-ES)` - `Farsi - Iran (fa-IR)` - `Finnish - Finland (fi-FI)` - `French - France (fr-FR)` - `Hebrew - Israel (he-IL)` - `Hungarian - Hungary (hu-HU)` - `Armenian - Armenia (hy-AM)` - `Indonesian - Indonesia (id-ID)` - `Icelandic - Iceland (is-IS)` - `Italian - Italy (it-IT)` - `Japanese - Japan (ja-JP)` - `Javanese - Indonesia (jv-ID)` - `Georgian - Georgia (ka-GE)` - `Khmer - Cambodia (km-KH)` - `Korean - Korea (ko-KR)` - `Latvian - Latvia (lv-LV)` - `Mongolian - Mongolia (mn-MN)` - `Malay - Malaysia (ms-MY)` - `Burmese - Myanmar (my-MM)` - `Norwegian - Norway (nb-NO)` - `Dutch - Netherlands (nl-NL)` - `Polish - Poland (pl-PL)` - `Portuguese - Portugal (pt-PT)` - `Romanian - Romania (ro-RO)` - `Russian - Russia (ru-RU)` - `Slovanian - Slovania (sl-SL)` - `Albanian - Albania (sq-AL)` - `Swedish - Sweden (sv-SE)` - `Swahili - Kenya (sw-KE)` - `Hindi - India (hi-IN)` - `Kannada - India (kn-IN)` - `Malayalam - India (ml-IN)` - `Tamil - India (ta-IN)` - `Telugu - India (te-IN)` - `Thai - Thailand (th-TH)` - `Tagalog - Philippines (tl-PH)` - `Turkish - Turkey (tr-TR)` - `Urdu - Pakistan (ur-PK)` - `Vietnamese - Vietnam (vi-VN)` - `Welsh - United Kingdom (cy-GB)` ## Evaluation results ```plain precision recall f1-score support af-ZA 0.9821 0.9805 0.9813 2974 am-ET 1.0000 1.0000 1.0000 2974 ar-SA 0.9809 0.9822 0.9815 2974 az-AZ 0.9946 0.9845 0.9895 2974 bn-BD 0.9997 0.9990 0.9993 2974 cy-GB 0.9970 0.9929 0.9949 2974 da-DK 0.9575 0.9617 0.9596 2974 de-DE 0.9906 0.9909 0.9908 2974 el-GR 0.9997 0.9973 0.9985 2974 en-US 0.9712 0.9866 0.9788 2974 es-ES 0.9825 0.9842 0.9834 2974 fa-IR 0.9940 0.9973 0.9956 2974 fi-FI 0.9943 0.9946 0.9945 2974 fr-FR 0.9963 0.9923 0.9943 2974 he-IL 1.0000 0.9997 0.9998 2974 hi-IN 1.0000 0.9980 0.9990 2974 hu-HU 0.9983 0.9950 0.9966 2974 hy-AM 1.0000 0.9993 0.9997 2974 id-ID 0.9319 0.9291 0.9305 2974 is-IS 0.9966 0.9943 0.9955 2974 it-IT 0.9698 0.9926 0.9811 2974 ja-JP 0.9987 0.9963 0.9975 2974 jv-ID 0.9628 0.9744 0.9686 2974 ka-GE 0.9993 0.9997 0.9995 2974 km-KH 0.9867 0.9963 0.9915 2974 kn-IN 1.0000 0.9993 0.9997 2974 ko-KR 0.9917 0.9997 0.9956 2974 lv-LV 0.9990 0.9950 0.9970 2974 ml-IN 0.9997 0.9997 0.9997 2974 mn-MN 0.9987 0.9966 0.9976 2974 ms-MY 0.9359 0.9418 0.9388 2974 my-MM 1.0000 0.9993 0.9997 2974 nb-NO 0.9600 0.9533 0.9566 2974 nl-NL 0.9850 0.9748 0.9799 2974 pl-PL 0.9946 0.9923 0.9934 2974 pt-PT 0.9885 0.9798 0.9841 2974 ro-RO 0.9919 0.9916 0.9918 2974 ru-RU 0.9976 0.9983 0.9980 2974 sl-SL 0.9956 0.9939 0.9948 2974 sq-AL 0.9936 0.9896 0.9916 2974 sv-SE 0.9902 0.9842 0.9872 2974 sw-KE 0.9867 0.9953 0.9910 2974 ta-IN 1.0000 1.0000 1.0000 2974 te-IN 1.0000 0.9997 0.9998 2974 th-TH 1.0000 0.9983 0.9992 2974 tl-PH 0.9929 0.9899 0.9914 2974 tr-TR 0.9869 0.9872 0.9871 2974 ur-PK 0.9983 0.9929 0.9956 2974 vi-VN 0.9993 0.9973 0.9983 2974 zh-CN 0.9812 0.9832 0.9822 2974 zh-TW 0.9832 0.9815 0.9823 2974 accuracy 0.9889 151674 macro avg 0.9889 0.9889 0.9889 151674 weighted avg 0.9889 0.9889 0.9889 151674 ``` Keywords : language identification ; language identification ; multilingual ; classification

timm/cspdarknet53.ra_in1k

timm

timm

image-classification

--- license: apache-2.0 library_name: timm tags: - image-classification - timm --- # Model card for cspdarknet53.ra_in1k A CSP-DarkNet (Cross-Stage-Partial) image classification model. Trained on ImageNet-1k in `timm` using recipe template described below. Recipe details: * RandAugment `RA` recipe. Inspired by and evolved from EfficientNet RandAugment recipes. Published as `B` recipe in [ResNet Strikes Back](https://arxiv.org/abs/2110.00476). * RMSProp (TF 1.0 behaviour) optimizer, EMA weight averaging * Step (exponential decay w/ staircase) LR schedule with warmup ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 27.6 - GMACs: 6.6 - Activations (M): 16.8 - Image size: 256 x 256 - **Papers:** - CSPNet: A New Backbone that can Enhance Learning Capability of CNN: https://arxiv.org/abs/1911.11929 - YOLOv3: An Incremental Improvement: https://arxiv.org/abs/1804.02767 - ResNet strikes back: An improved training procedure in timm: https://arxiv.org/abs/2110.00476 - **Original:** https://github.com/huggingface/pytorch-image-models ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('cspdarknet53.ra_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'cspdarknet53.ra_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 32, 256, 256]) # torch.Size([1, 64, 128, 128]) # torch.Size([1, 128, 64, 64]) # torch.Size([1, 256, 32, 32]) # torch.Size([1, 512, 16, 16]) # torch.Size([1, 1024, 8, 8]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'cspdarknet53.ra_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1024, 8, 8) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{Wang2019CSPNetAN, title={CSPNet: A New Backbone that can Enhance Learning Capability of CNN}, author={Chien-Yao Wang and Hong-Yuan Mark Liao and I-Hau Yeh and Yueh-Hua Wu and Ping-Yang Chen and Jun-Wei Hsieh}, journal={2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)}, year={2019}, pages={1571-1580} } ``` ```bibtex @article{Redmon2018YOLOv3AI, title={YOLOv3: An Incremental Improvement}, author={Joseph Redmon and Ali Farhadi}, journal={ArXiv}, year={2018}, volume={abs/1804.02767} } ``` ```bibtex @inproceedings{wightman2021resnet, title={ResNet strikes back: An improved training procedure in timm}, author={Wightman, Ross and Touvron, Hugo and Jegou, Herve}, booktitle={NeurIPS 2021 Workshop on ImageNet: Past, Present, and Future} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

microsoft/llmlingua-2-bert-base-multilingual-cased-meetingbank

microsoft

transformers

token-classification

--- license: apache-2.0 --- # LLMLingua-2-Bert-base-Multilingual-Cased-MeetingBank This model was introduced in the paper [**LLMLingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression** (Pan et al, 2024)](https://arxiv.org/abs/2403.12968). It is a [BERT multilingual base model (cased)](https://huggingface.co/google-bert/bert-base-multilingual-cased) finetuned to perform token classification for task agnostic prompt compression. The probability `$p_{preserve}$` of each token `$x_i$` is used as the metric for compression. This model is trained on [an extractive text compression dataset(will public)]() constructed with the methodology proposed in the [**LLMLingua-2**](https://arxiv.org/abs/2403.12968), using training examples from [MeetingBank (Hu et al, 2023)](https://meetingbank.github.io/) as the seed data. For more details, please check the project page of [LLMLingua-2](https://llmlingua.com/llmlingua2.html) and [LLMLingua Series](https://llmlingua.com/). ## Usage ```python from llmlingua import PromptCompressor compressor = PromptCompressor( model_name="microsoft/llmlingua-2-bert-base-multilingual-cased-meetingbank", use_llmlingua2=True ) original_prompt = """John: So, um, I've been thinking about the project, you know, and I believe we need to, uh, make some changes. I mean, we want the project to succeed, right? So, like, I think we should consider maybe revising the timeline. Sarah: I totally agree, John. I mean, we have to be realistic, you know. The timeline is, like, too tight. You know what I mean? We should definitely extend it. """ results = compressor.compress_prompt_llmlingua2( original_prompt, rate=0.6, force_tokens=['\n', '.', '!', '?', ','], chunk_end_tokens=['.', '\n'], return_word_label=True, drop_consecutive=True ) print(results.keys()) print(f"Compressed prompt: {results['compressed_prompt']}") print(f"Original tokens: {results['origin_tokens']}") print(f"Compressed tokens: {results['compressed_tokens']}") print(f"Compression rate: {results['rate']}") # get the annotated results over the original prompt word_sep = "\t\t|\t\t" label_sep = " " lines = results["fn_labeled_original_prompt"].split(word_sep) annotated_results = [] for line in lines: word, label = line.split(label_sep) annotated_results.append((word, '+') if label == '1' else (word, '-')) # list of tuples: (word, label) print("Annotated results:") for word, label in annotated_results[:10]: print(f"{word} {label}") ``` ## Citation ``` @article{wu2024llmlingua2, title = "{LLML}ingua-2: Data Distillation for Efficient and Faithful Task-Agnostic Prompt Compression", author = "Zhuoshi Pan and Qianhui Wu and Huiqiang Jiang and Menglin Xia and Xufang Luo and Jue Zhang and Qingwei Lin and Victor Ruhle and Yuqing Yang and Chin-Yew Lin and H. Vicky Zhao and Lili Qiu and Dongmei Zhang", url = "https://arxiv.org/abs/2403.12968", journal = "ArXiv preprint", volume = "abs/2403.12968", year = "2024", } ```

timm/eva02_base_patch14_224.mim_in22k

timm

timm

image-feature-extraction

--- license: mit library_name: timm tags: - image-feature-extraction - timm --- # Model card for eva02_base_patch14_224.mim_in22k An EVA02 feature / representation model. Pretrained on ImageNet-22k with masked image modeling (using EVA-CLIP as a MIM teacher) by paper authors. EVA-02 models are vision transformers with mean pooling, SwiGLU, Rotary Position Embeddings (ROPE), and extra LN in MLP (for Base & Large). NOTE: `timm` checkpoints are float32 for consistency with other models. Original checkpoints are float16 or bfloat16 in some cases, see originals if that's preferred. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 85.8 - GMACs: 23.2 - Activations (M): 36.6 - Image size: 224 x 224 - **Papers:** - EVA-02: A Visual Representation for Neon Genesis: https://arxiv.org/abs/2303.11331 - EVA-CLIP: Improved Training Techniques for CLIP at Scale: https://arxiv.org/abs/2303.15389 - **Original:** - https://github.com/baaivision/EVA - https://huggingface.co/Yuxin-CV/EVA-02 - **Pretrain Dataset:** ImageNet-22k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('eva02_base_patch14_224.mim_in22k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'eva02_base_patch14_224.mim_in22k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 257, 768) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). |model |top1 |top5 |param_count|img_size| |-----------------------------------------------|------|------|-----------|--------| |eva02_large_patch14_448.mim_m38m_ft_in22k_in1k |90.054|99.042|305.08 |448 | |eva02_large_patch14_448.mim_in22k_ft_in22k_in1k|89.946|99.01 |305.08 |448 | |eva_giant_patch14_560.m30m_ft_in22k_in1k |89.792|98.992|1014.45 |560 | |eva02_large_patch14_448.mim_in22k_ft_in1k |89.626|98.954|305.08 |448 | |eva02_large_patch14_448.mim_m38m_ft_in1k |89.57 |98.918|305.08 |448 | |eva_giant_patch14_336.m30m_ft_in22k_in1k |89.56 |98.956|1013.01 |336 | |eva_giant_patch14_336.clip_ft_in1k |89.466|98.82 |1013.01 |336 | |eva_large_patch14_336.in22k_ft_in22k_in1k |89.214|98.854|304.53 |336 | |eva_giant_patch14_224.clip_ft_in1k |88.882|98.678|1012.56 |224 | |eva02_base_patch14_448.mim_in22k_ft_in22k_in1k |88.692|98.722|87.12 |448 | |eva_large_patch14_336.in22k_ft_in1k |88.652|98.722|304.53 |336 | |eva_large_patch14_196.in22k_ft_in22k_in1k |88.592|98.656|304.14 |196 | |eva02_base_patch14_448.mim_in22k_ft_in1k |88.23 |98.564|87.12 |448 | |eva_large_patch14_196.in22k_ft_in1k |87.934|98.504|304.14 |196 | |eva02_small_patch14_336.mim_in22k_ft_in1k |85.74 |97.614|22.13 |336 | |eva02_tiny_patch14_336.mim_in22k_ft_in1k |80.658|95.524|5.76 |336 | ## Citation ```bibtex @article{EVA02, title={EVA-02: A Visual Representation for Neon Genesis}, author={Fang, Yuxin and Sun, Quan and Wang, Xinggang and Huang, Tiejun and Wang, Xinlong and Cao, Yue}, journal={arXiv preprint arXiv:2303.11331}, year={2023} } ``` ```bibtex @article{EVA-CLIP, title={EVA-02: A Visual Representation for Neon Genesis}, author={Sun, Quan and Fang, Yuxin and Wu, Ledell and Wang, Xinlong and Cao, Yue}, journal={arXiv preprint arXiv:2303.15389}, year={2023} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

setu4993/LaBSE

setu4993

transformers

sentence-similarity

--- pipeline_tag: sentence-similarity language: - af - am - ar - as - az - be - bg - bn - bo - bs - ca - ceb - co - cs - cy - da - de - el - en - eo - es - et - eu - fa - fi - fr - fy - ga - gd - gl - gu - ha - haw - he - hi - hmn - hr - ht - hu - hy - id - ig - is - it - ja - jv - ka - kk - km - kn - ko - ku - ky - la - lb - lo - lt - lv - mg - mi - mk - ml - mn - mr - ms - mt - my - ne - nl - no - ny - or - pa - pl - pt - ro - ru - rw - si - sk - sl - sm - sn - so - sq - sr - st - su - sv - sw - ta - te - tg - th - tk - tl - tr - tt - ug - uk - ur - uz - vi - wo - xh - yi - yo - zh - zu tags: - bert - sentence_embedding - multilingual - google - sentence-similarity license: apache-2.0 datasets: - CommonCrawl - Wikipedia --- # LaBSE ## Model description Language-agnostic BERT Sentence Encoder (LaBSE) is a BERT-based model trained for sentence embedding for 109 languages. The pre-training process combines masked language modeling with translation language modeling. The model is useful for getting multilingual sentence embeddings and for bi-text retrieval. - Model: [HuggingFace's model hub](https://huggingface.co/setu4993/LaBSE). - Paper: [arXiv](https://arxiv.org/abs/2007.01852). - Original model: [TensorFlow Hub](https://tfhub.dev/google/LaBSE/2). - Blog post: [Google AI Blog](https://ai.googleblog.com/2020/08/language-agnostic-bert-sentence.html). - Conversion from TensorFlow to PyTorch: [GitHub](https://github.com/setu4993/convert-labse-tf-pt). This is migrated from the v2 model on the TF Hub, which uses dict-based input. The embeddings produced by both the versions of the model are [equivalent](https://github.com/setu4993/convert-labse-tf-pt/blob/ec3a019159a54ed6493181a64486c2808c01f216/tests/test_conversion.py#L31). ## Usage Using the model: ```python import torch from transformers import BertModel, BertTokenizerFast tokenizer = BertTokenizerFast.from_pretrained("setu4993/LaBSE") model = BertModel.from_pretrained("setu4993/LaBSE") model = model.eval() english_sentences = [ "dog", "Puppies are nice.", "I enjoy taking long walks along the beach with my dog.", ] english_inputs = tokenizer(english_sentences, return_tensors="pt", padding=True) with torch.no_grad(): english_outputs = model(**english_inputs) ``` To get the sentence embeddings, use the pooler output: ```python english_embeddings = english_outputs.pooler_output ``` Output for other languages: ```python italian_sentences = [ "cane", "I cuccioli sono carini.", "Mi piace fare lunghe passeggiate lungo la spiaggia con il mio cane.", ] japanese_sentences = ["犬", "子犬はいいです", "私は犬と一緒にビーチを散歩するのが好きです"] italian_inputs = tokenizer(italian_sentences, return_tensors="pt", padding=True) japanese_inputs = tokenizer(japanese_sentences, return_tensors="pt", padding=True) with torch.no_grad(): italian_outputs = model(**italian_inputs) japanese_outputs = model(**japanese_inputs) italian_embeddings = italian_outputs.pooler_output japanese_embeddings = japanese_outputs.pooler_output ``` For similarity between sentences, an L2-norm is recommended before calculating the similarity: ```python import torch.nn.functional as F def similarity(embeddings_1, embeddings_2): normalized_embeddings_1 = F.normalize(embeddings_1, p=2) normalized_embeddings_2 = F.normalize(embeddings_2, p=2) return torch.matmul( normalized_embeddings_1, normalized_embeddings_2.transpose(0, 1) ) print(similarity(english_embeddings, italian_embeddings)) print(similarity(english_embeddings, japanese_embeddings)) print(similarity(italian_embeddings, japanese_embeddings)) ``` ## Details Details about data, training, evaluation and performance metrics are available in the [original paper](https://arxiv.org/abs/2007.01852). ### BibTeX entry and citation info ```bibtex @misc{feng2020languageagnostic, title={Language-agnostic BERT Sentence Embedding}, author={Fangxiaoyu Feng and Yinfei Yang and Daniel Cer and Naveen Arivazhagan and Wei Wang}, year={2020}, eprint={2007.01852}, archivePrefix={arXiv}, primaryClass={cs.CL} } ```

Qwen/Qwen2-72B-Instruct-GPTQ-Int4

Qwen

transformers

text-generation

--- license: other license_name: tongyi-qianwen license_link: https://huggingface.co/Qwen/Qwen2-72B-Instruct-GPTQ-Int4/blob/main/LICENSE language: - en pipeline_tag: text-generation tags: - chat --- # Qwen2-72B-Instruct-GPTQ-Int4 ## Introduction Qwen2 is the new series of Qwen large language models. For Qwen2, we release a number of base language models and instruction-tuned language models ranging from 0.5 to 72 billion parameters, including a Mixture-of-Experts model. This repo contains the instruction-tuned 72B Qwen2 model. Compared with the state-of-the-art opensource language models, including the previous released Qwen1.5, Qwen2 has generally surpassed most opensource models and demonstrated competitiveness against proprietary models across a series of benchmarks targeting for language understanding, language generation, multilingual capability, coding, mathematics, reasoning, etc. Qwen2-72B-Instruct-GPTQ-Int4 supports a context length of up to 131,072 tokens, enabling the processing of extensive inputs. Please refer to [this section](#processing-long-texts) for detailed instructions on how to deploy Qwen2 for handling long texts. For more details, please refer to our [blog](https://qwenlm.github.io/blog/qwen2/), [GitHub](https://github.com/QwenLM/Qwen2), and [Documentation](https://qwen.readthedocs.io/en/latest/). <br> ## Model Details Qwen2 is a language model series including decoder language models of different model sizes. For each size, we release the base language model and the aligned chat model. It is based on the Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, etc. Additionally, we have an improved tokenizer adaptive to multiple natural languages and codes. ## Training details We pretrained the models with a large amount of data, and we post-trained the models with both supervised finetuning and direct preference optimization. ## Requirements The code of Qwen2 has been in the latest Hugging face transformers and we advise you to install `transformers>=4.37.0`, or you might encounter the following error: ``` KeyError: 'qwen2' ``` ## Quickstart Here provides a code snippet with `apply_chat_template` to show you how to load the tokenizer and model and how to generate contents. ```python from transformers import AutoModelForCausalLM, AutoTokenizer device = "cuda" # the device to load the model onto model = AutoModelForCausalLM.from_pretrained( "Qwen/Qwen2-72B-Instruct-GPTQ-Int4", torch_dtype="auto", device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2-72B-Instruct-GPTQ-Int4") prompt = "Give me a short introduction to large language model." messages = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt} ] text = tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) model_inputs = tokenizer([text], return_tensors="pt").to(device) generated_ids = model.generate( model_inputs.input_ids, max_new_tokens=512 ) generated_ids = [ output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids) ] response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0] ``` ### Processing Long Texts To handle extensive inputs exceeding 32,768 tokens, we utilize [YARN](https://arxiv.org/abs/2309.00071), a technique for enhancing model length extrapolation, ensuring optimal performance on lengthy texts. For deployment, we recommend using vLLM. You can enable the long-context capabilities by following these steps: 1. **Install vLLM**: You can install vLLM by running the following command. ```bash pip install "vllm>=0.4.3" ``` Or you can install vLLM from [source](https://github.com/vllm-project/vllm/). 2. **Configure Model Settings**: After downloading the model weights, modify the `config.json` file by including the below snippet: ```json { "architectures": [ "Qwen2ForCausalLM" ], // ... "vocab_size": 152064, // adding the following snippets "rope_scaling": { "factor": 4.0, "original_max_position_embeddings": 32768, "type": "yarn" } } ``` This snippet enable YARN to support longer contexts. 3. **Model Deployment**: Utilize vLLM to deploy your model. For instance, you can set up an openAI-like server using the command: ```bash python -m vllm.entrypoints.openai.api_server --served-model-name Qwen2-72B-Instruct-GPTQ-Int4 --model path/to/weights ``` Then you can access the Chat API by: ```bash curl http://localhost:8000/v1/chat/completions \ -H "Content-Type: application/json" \ -d '{ "model": "Qwen2-72B-Instruct-GPTQ-Int4", "messages": [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": "Your Long Input Here."} ] }' ``` For further usage instructions of vLLM, please refer to our [Github](https://github.com/QwenLM/Qwen2). **Note**: Presently, vLLM only supports static YARN, which means the scaling factor remains constant regardless of input length, **potentially impacting performance on shorter texts**. We advise adding the `rope_scaling` configuration only when processing long contexts is required. ## Benchmark and Speed To compare the generation performance between bfloat16 (bf16) and quantized models such as GPTQ-Int8, GPTQ-Int4, and AWQ, please consult our [Benchmark of Quantized Models](https://qwen.readthedocs.io/en/latest/benchmark/quantization_benchmark.html). This benchmark provides insights into how different quantization techniques affect model performance. For those interested in understanding the inference speed and memory consumption when deploying these models with either ``transformer`` or ``vLLM``, we have compiled an extensive [Speed Benchmark](https://qwen.readthedocs.io/en/latest/benchmark/speed_benchmark.html). ## Citation If you find our work helpful, feel free to give us a cite. ``` @article{qwen2, title={Qwen2 Technical Report}, year={2024} } ```

microsoft/wavlm-base-plus-sv

microsoft

transformers

--- language: - en tags: - speech --- # WavLM-Base-Plus for Speaker Verification [Microsoft's WavLM](https://github.com/microsoft/unilm/tree/master/wavlm) The model was pretrained on 16kHz sampled speech audio with utterance and speaker contrastive loss. When using the model, make sure that your speech input is also sampled at 16kHz. **Note**: This model does not have a tokenizer as it was pretrained on audio alone. In order to use this model **speech recognition**, a tokenizer should be created and the model should be fine-tuned on labeled text data. Check out [this blog](https://huggingface.co/blog/fine-tune-wav2vec2-english) for more in-detail explanation of how to fine-tune the model. The model was pre-trained on: - 60,000 hours of [Libri-Light](https://arxiv.org/abs/1912.07875) - 10,000 hours of [GigaSpeech](https://arxiv.org/abs/2106.06909) - 24,000 hours of [VoxPopuli](https://arxiv.org/abs/2101.00390) [Paper: WavLM: Large-Scale Self-Supervised Pre-Training for Full Stack Speech Processing](https://arxiv.org/abs/2110.13900) Authors: Sanyuan Chen, Chengyi Wang, Zhengyang Chen, Yu Wu, Shujie Liu, Zhuo Chen, Jinyu Li, Naoyuki Kanda, Takuya Yoshioka, Xiong Xiao, Jian Wu, Long Zhou, Shuo Ren, Yanmin Qian, Yao Qian, Jian Wu, Michael Zeng, Furu Wei **Abstract** *Self-supervised learning (SSL) achieves great success in speech recognition, while limited exploration has been attempted for other speech processing tasks. As speech signal contains multi-faceted information including speaker identity, paralinguistics, spoken content, etc., learning universal representations for all speech tasks is challenging. In this paper, we propose a new pre-trained model, WavLM, to solve full-stack downstream speech tasks. WavLM is built based on the HuBERT framework, with an emphasis on both spoken content modeling and speaker identity preservation. We first equip the Transformer structure with gated relative position bias to improve its capability on recognition tasks. For better speaker discrimination, we propose an utterance mixing training strategy, where additional overlapped utterances are created unsupervisely and incorporated during model training. Lastly, we scale up the training dataset from 60k hours to 94k hours. WavLM Large achieves state-of-the-art performance on the SUPERB benchmark, and brings significant improvements for various speech processing tasks on their representative benchmarks.* The original model can be found under https://github.com/microsoft/unilm/tree/master/wavlm. # Fine-tuning details The model is fine-tuned on the [VoxCeleb1 dataset](https://www.robots.ox.ac.uk/~vgg/data/voxceleb/vox1.html) using an X-Vector head with an Additive Margin Softmax loss [X-Vectors: Robust DNN Embeddings for Speaker Recognition](https://www.danielpovey.com/files/2018_icassp_xvectors.pdf) # Usage ## Speaker Verification ```python from transformers import Wav2Vec2FeatureExtractor, WavLMForXVector from datasets import load_dataset import torch dataset = load_dataset("hf-internal-testing/librispeech_asr_demo", "clean", split="validation") feature_extractor = Wav2Vec2FeatureExtractor.from_pretrained('microsoft/wavlm-base-plus-sv') model = WavLMForXVector.from_pretrained('microsoft/wavlm-base-plus-sv') # audio files are decoded on the fly audio = [x["array"] for x in dataset[:2]["audio"]] inputs = feature_extractor(audio, padding=True, return_tensors="pt") embeddings = model(**inputs).embeddings embeddings = torch.nn.functional.normalize(embeddings, dim=-1).cpu() # the resulting embeddings can be used for cosine similarity-based retrieval cosine_sim = torch.nn.CosineSimilarity(dim=-1) similarity = cosine_sim(embeddings[0], embeddings[1]) threshold = 0.86 # the optimal threshold is dataset-dependent if similarity < threshold: print("Speakers are not the same!") ``` # License The official license can be found [here](https://github.com/microsoft/UniSpeech/blob/main/LICENSE) 

pkshatech/GLuCoSE-base-ja

pkshatech

sentence-transformers

sentence-similarity

--- pipeline_tag: sentence-similarity language: ja license: apache-2.0 tags: - transformers - sentence-similarity - feature-extraction - sentence-transformers inference: false datasets: - mc4 - clips/mqa - shunk031/JGLUE - paws-x - MoritzLaurer/multilingual-NLI-26lang-2mil7 - castorini/mr-tydi - hpprc/jsick --- # GLuCoSE (General Luke-based Contrastive Sentence Embedding)-base-Japanese [日本語のREADME/Japanese README](https://huggingface.co/pkshatech/GLuCoSE-base-ja/blob/main/README_JA.md) GLuCoSE (General LUke-based COntrastive Sentence Embedding, "glucose") is a Japanese text embedding model based on [LUKE](https://github.com/studio-ousia/luke). In order to create a general-purpose, user-friendly Japanese text embedding model, GLuCoSE has been trained on a mix of web data and various datasets associated with natural language inference and search. This model is not only suitable for sentence vector similarity tasks but also for semantic search tasks. - Maximum token count: 512 - Output dimension: 768 - Pooling: mean pooling - Supported language: Japanese ## Usage You can use this model easily with [sentence-transformers](https://www.SBERT.net). First, install sentence-transformers with pip as follows: ``` pip install -U sentence-transformers ``` You can load the model and convert sentences into dense vectors as shown below: ```python from sentence_transformers import SentenceTransformer sentences = [ "PKSHA Technologyは機械学習/深層学習技術に関わるアルゴリズムソリューションを展開している。", "この深層学習モデルはPKSHA Technologyによって学習され、公開された。", "広目天は、仏教における四天王の一尊であり、サンスクリット語の「種々の眼をした者」を名前の由来とする。", ] model = SentenceTransformer('pkshatech/GLuCoSE-base-ja') embeddings = model.encode(sentences) print(embeddings) ``` Since the loss function used during training is cosine similarity, we recommend using cosine similarity for downstream tasks. This text embedding model can also be used in LangChain. Please refer to [this page](https://python.langchain.com/docs/modules/data_connection/text_embedding/integrations/sentence_transformers) for more information. ## Resources Used The following resources were used to train this model. ### Pre-trained model - [studio-ousia/luke-japanese-base-lite](https://huggingface.co/studio-ousia/luke-japanese-base-lite) ### Datasets - [mC4](https://huggingface.co/datasets/mc4) - [MQA](https://huggingface.co/datasets/clips/mqa) - [JNLI](https://github.com/yahoojapan/JGLUE) - [JSNLI](https://nlp.ist.i.kyoto-u.ac.jp/?%E6%97%A5%E6%9C%AC%E8%AA%9ESNLI%28JSNLI%29%E3%83%87%E3%83%BC%E3%82%BF%E3%82%BB%E3%83%83%E3%83%88) - [PAWS-X](https://huggingface.co/datasets/paws-x) - [JSeM](https://github.com/DaisukeBekki/JSeM) - [MoritzLaurer/multilingual-NLI-26lang-2mil7](https://huggingface.co/datasets/MoritzLaurer/multilingual-NLI-26lang-2mil7) - [MultiNLI](https://huggingface.co/datasets/multi_nli) - [WANLI](https://huggingface.co/datasets/alisawuffles/WANLI) - [FeverNLI](https://github.com/easonnie/combine-FEVER-NSMN/blob/master/other_resources/nli_fever.md) - [LingNLI](https://arxiv.org/pdf/2104.07179.pdf) - [JSICK](https://github.com/verypluming/JSICK) - [Mr.Tidy](https://huggingface.co/datasets/castorini/mr-tydi) - [JSTS](https://github.com/yahoojapan/JGLUE) (used for validation) [^1] ## Benchmarks ### Semantic Similarity Calculation ([JSTS](https://github.com/yahoojapan/JGLUE) dev set) Evaluation by Spearman's correlation coefficient and Pearson's correlation coefficient. | Model | Spearman | Pearson | | --- | --- | --- | | [text-embedding-ada-002](https://platform.openai.com/docs/guides/embeddings) |0.837[^2] | 0.790[^2] | | [pkshatech/simcse-ja-bert-base-clcmlp](https://huggingface.co/pkshatech/simcse-ja-bert-base-clcmlp)[^3] | 0.850 | 0.801 | | pkshatech/GLuCoSE-base-ja | **0.864** | **0.818** | ### Zero-shot Search ([AIO3](https://sites.google.com/view/project-aio/competition3?authuser=0) dev set) Evaluation by top-k retrieval accuracy[^4] (the fraction of questions that have a correct answer in the top-k retrieved documents at least once.) | Model | Top-1 | Top-5 | Top-10 | Top-50 | | --- | --- | --- | --- | --- | | [text-embedding-ada-002](https://platform.openai.com/docs/guides/embeddings) | 33.50 | 57.80 | 65.10 | 76.60 | | [pkshatech/simcse-ja-bert-base-clcmlp](https://huggingface.co/pkshatech/simcse-ja-bert-base-clcmlp)[^3] | 30.60 | 54.50 | 62.50 | 76.70 | | pkshatech/GLuCoSE-base-ja | **36.10** | **59.40** | **66.40** | **78.30** | # Authors [Akihiko Fukuchi](https://huggingface.co/akiFQC), [Yuichiro Hoshino](https://huggingface.co/Yuichiroh), [Yotarow Watanabe](https://huggingface.co/yotarow) ## License This model is published under the [Apache License, Version 2.0](https://www.apache.org/licenses/LICENSE-2.0). [^1]: When we trained this model, the test data of JGLUE was not released, so we used the dev set of JGLUE as a private evaluation data. Therefore, we selected the checkpoint on the train set of JGLUE insted of its dev set. [^2]: https://qiita.com/akeyhero/items/ce371bfed64399027c23 [^3]: This is the model we have released before. [^4]: For more details, please refer to https://arxiv.org/pdf/2004.04906.pdf.

mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF

mradermacher

transformers

--- base_model: cloudyu/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO language: - en library_name: transformers license: apache-2.0 quantized_by: mradermacher tags: - yi - moe --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: --> static quants of https://huggingface.co/cloudyu/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO <!-- provided-files --> weighted/imatrix quants are available at https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-i1-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q2_K.gguf) | Q2_K | 22.5 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.IQ3_XS.gguf) | IQ3_XS | 25.1 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q3_K_S.gguf) | Q3_K_S | 26.4 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.IQ3_S.gguf) | IQ3_S | 26.5 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.IQ3_M.gguf) | IQ3_M | 27.2 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q3_K_M.gguf) | Q3_K_M | 29.3 | lower quality | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q3_K_L.gguf) | Q3_K_L | 31.9 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.IQ4_XS.gguf) | IQ4_XS | 32.9 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q4_K_S.gguf) | Q4_K_S | 34.7 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q4_K_M.gguf) | Q4_K_M | 36.8 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q5_K_S.gguf) | Q5_K_S | 42.0 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q5_K_M.gguf) | Q5_K_M | 43.2 | | | [GGUF](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q6_K.gguf) | Q6_K | 50.0 | very good quality | | [PART 1](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q8_0.gguf.part1of2) [PART 2](https://huggingface.co/mradermacher/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO-GGUF/resolve/main/TomGrc_FusionNet_34Bx2_MoE_v0.1_full_linear_DPO.Q8_0.gguf.part2of2) | Q8_0 | 64.7 | fast, best quality | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):  And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. <!-- end -->

marcobombieri/surgicberta

marcobombieri

transformers

fill-mask

--- license: cc-by-nc-nd-4.0 widget: - text: "hemithyroidectomy is the removal of half of the <mask> gland." example_title: "Example 1" - text: "Li-Fraumeni <mask> is a hereditary tumor with autosomal dominant inheritance." example_title: "Example 2" - text: "The fascia in front of the pancreas was cut to the spleen direction to <mask> the splenic artery and vein." example_title: "Example 3" --- # SurgicBERTa SurgicBERTa is a language model based on RoBERTa-base (Liu et al., 2019) architecture. We adapted RoBERTa-base to different **surgical textbooks and academic papers** via continued pretraining. This amount to about 7 million words and 300k surgical sentences. We used the full text of the books and papers in training, not just abstracts. Specific details of the adaptive pretraining procedure and evaluation tasks can be found in the paper below cited. # Citation If using this model, please cite the following paper: <em> <br /> <br /> @article{bombieri_et_al_SurgicBERTa_2023, <br /> title = {Surgicberta: a pre-trained language model for procedural surgical language}, <br /> journal = {International Journal of Data Science and Analytics}, <br /> year = {2023}, <br /> doi = { https://doi.org/10.1007/s41060-023-00433-5 }, <br /> url = { https://link.springer.com/article/10.1007/s41060-023-00433-5 }, <br /> author = {Marco Bombieri and Marco Rospocher and Simone Paolo Ponzetto and Paolo Fiorini}, <br /> } </em> If using this model for Semantic Role Labeling, please cite also the following paper: <em> <br /> <br /> @article{bombieri_et_al_surgical_srl_2023, <br /> title = {Machine understanding surgical actions from intervention procedure textbooks}, <br /> journal = {Computers in Biology and Medicine}, <br /> volume = {152}, <br /> pages = {106415}, <br /> year = {2023}, <br /> issn = {0010-4825}, <br /> doi = {https://doi.org/10.1016/j.compbiomed.2022.106415}, <br /> url = {https://www.sciencedirect.com/science/article/pii/S0010482522011234}, <br /> author = {Marco Bombieri and Marco Rospocher and Simone Paolo Ponzetto and Paolo Fiorini}, <br /> keywords = {Semantic role labeling, Surgical data science, Procedural knowledge, Information extraction, Natural language processing} <br /> } </em>

SG161222/RealVisXL_V3.0

SG161222

diffusers

text-to-image

--- license: openrail++ --- <b>This model is available on <a href="https://www.mage.space/">Mage.Space</a> (main sponsor)</b><br> <b>You can support me directly on Boosty - https://boosty.to/sg_161222</b><br> <b>It's important! Read it!</b><br> The model is still in the training phase. This is not the final version and may contain artifacts and perform poorly in some cases.<br> The model is aimed at photorealism. Can produce sfw and nsfw images of decent quality.<br> CivitAI Page: https://civitai.com/models/139562/realvisxl-v30-turbo<br> <b>Recommended Negative Prompt:</b><br> (worst quality, low quality, illustration, 3d, 2d, painting, cartoons, sketch), open mouth<br> <b>or another negative prompt</b><br> <b>Recommended Generation Parameters:</b><br> Sampling Steps: 15-30<br> Sampling Method: DPM++ SDE Karras<br> <b>Recommended Hires Fix Parameters:</b><br> Hires steps: 10+<br> Upscaler: 4x-UltraSharp upscaler / or another<br> Denoising strength: 0.1 - 0.5<br> Upscale by: 1.1-2.0<br>

mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF

mradermacher

transformers

--- base_model: 01-ai/Yi-1.5-34B-Chat-16K language: - en library_name: transformers license: apache-2.0 quantized_by: mradermacher --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: nicoboss --> weighted/imatrix quants of https://huggingface.co/01-ai/Yi-1.5-34B-Chat-16K <!-- provided-files --> static quants are available at https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ1_S.gguf) | i1-IQ1_S | 7.6 | for the desperate | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ1_M.gguf) | i1-IQ1_M | 8.3 | mostly desperate | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 9.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ2_XS.gguf) | i1-IQ2_XS | 10.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ2_S.gguf) | i1-IQ2_S | 11.0 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ2_M.gguf) | i1-IQ2_M | 11.9 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q2_K.gguf) | i1-Q2_K | 12.9 | IQ3_XXS probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 13.4 | lower quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ3_XS.gguf) | i1-IQ3_XS | 14.3 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q3_K_S.gguf) | i1-Q3_K_S | 15.1 | IQ3_XS probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ3_S.gguf) | i1-IQ3_S | 15.1 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ3_M.gguf) | i1-IQ3_M | 15.7 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q3_K_M.gguf) | i1-Q3_K_M | 16.8 | IQ3_S probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q3_K_L.gguf) | i1-Q3_K_L | 18.2 | IQ3_M probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-IQ4_XS.gguf) | i1-IQ4_XS | 18.6 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q4_0.gguf) | i1-Q4_0 | 19.6 | fast, low quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q4_K_S.gguf) | i1-Q4_K_S | 19.7 | optimal size/speed/quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q4_K_M.gguf) | i1-Q4_K_M | 20.8 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q5_K_S.gguf) | i1-Q5_K_S | 23.8 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q5_K_M.gguf) | i1-Q5_K_M | 24.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-Chat-16K-i1-GGUF/resolve/main/Yi-1.5-34B-Chat-16K.i1-Q6_K.gguf) | i1-Q6_K | 28.3 | practically like static Q6_K | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):  And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his hardware for calculating the imatrix for these quants. <!-- end -->

EleutherAI/pythia-1b-deduped

EleutherAI

transformers

text-generation

--- language: - en tags: - pytorch - causal-lm - pythia license: apache-2.0 datasets: - EleutherAI/the_pile_deduplicated --- The *Pythia Scaling Suite* is a collection of models developed to facilitate interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf). It contains two sets of eight models of sizes 70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two models: one trained on the Pile, and one trained on the Pile after the dataset has been globally deduplicated. All 8 model sizes are trained on the exact same data, in the exact same order. We also provide 154 intermediate checkpoints per model, hosted on Hugging Face as branches. The Pythia model suite was designed to promote scientific research on large language models, especially interpretability research. Despite not centering downstream performance as a design goal, we find the models <a href="#evaluations">match or exceed</a> the performance of similar and same-sized models, such as those in the OPT and GPT-Neo suites. <details> <summary style="font-weight:600">Details on previous early release and naming convention.</summary> Previously, we released an early version of the Pythia suite to the public. However, we decided to retrain the model suite to address a few hyperparameter discrepancies. This model card <a href="#changelog">lists the changes</a>; see appendix B in the Pythia paper for further discussion. We found no difference in benchmark performance between the two Pythia versions. The old models are [still available](https://huggingface.co/models?other=pythia_v0), but we suggest the retrained suite if you are just starting to use Pythia.<br> **This is the current release.** Please note that all models in the *Pythia* suite were renamed in January 2023. For clarity, a <a href="#naming-convention-and-parameter-count">table comparing the old and new names</a> is provided in this model card, together with exact parameter counts. </details> <br> # Pythia-1B-deduped ## Model Details - Developed by: [EleutherAI](http://eleuther.ai) - Model type: Transformer-based Language Model - Language: English - Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia) for training procedure, config files, and details on how to use. [See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation details. - Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox) - License: Apache 2.0 - Contact: to ask questions about this model, join the [EleutherAI Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`. Please read the existing *Pythia* documentation before asking about it in the EleutherAI Discord. For general correspondence: [contact@eleuther. ai](mailto:contact@eleuther.ai). <figure> | Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models | | -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: | | 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10^-3 | — | | 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10^-4 | GPT-Neo 125M, OPT-125M | | 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10^-4 | OPT-350M | | 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10^-4 | — | | 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10^-4 | GPT-Neo 1.3B, OPT-1.3B | | 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10^-4 | GPT-Neo 2.7B, OPT-2.7B | | 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10^-4 | OPT-6.7B | | 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10^-4 | — | <figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and non-deduped models of a given size have the same hyperparameters. “Equivalent” models have <b>exactly</b> the same architecture, and the same number of non-embedding parameters.</figcaption> </figure> ## Uses and Limitations ### Intended Use The primary intended use of Pythia is research on the behavior, functionality, and limitations of large language models. This suite is intended to provide a controlled setting for performing scientific experiments. We also provide 154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints `step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to `step143000`. These checkpoints are hosted on Hugging Face as branches. Note that branch `143000` corresponds exactly to the model checkpoint on the `main` branch of each model. You may also further fine-tune and adapt Pythia-1B-deduped for deployment, as long as your use is in accordance with the Apache 2.0 license. Pythia models work with the Hugging Face [Transformers Library](https://huggingface.co/docs/transformers/index). If you decide to use pre-trained Pythia-1B-deduped as a basis for your fine-tuned model, please conduct your own risk and bias assessment. ### Out-of-scope use The Pythia Suite is **not** intended for deployment. It is not a in itself a product and cannot be used for human-facing interactions. For example, the model may generate harmful or offensive text. Please evaluate the risks associated with your particular use case. Pythia models are English-language only, and are not suitable for translation or generating text in other languages. Pythia-1B-deduped has not been fine-tuned for downstream contexts in which language models are commonly deployed, such as writing genre prose, or commercial chatbots. This means Pythia-1B-deduped will **not** respond to a given prompt the way a product like ChatGPT does. This is because, unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement Learning from Human Feedback (RLHF) to better “follow” human instructions. ### Limitations and biases The core functionality of a large language model is to take a string of text and predict the next token. The token used by the model need not produce the most “accurate” text. Never rely on Pythia-1B-deduped to produce factually accurate output. This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset known to contain profanity and texts that are lewd or otherwise offensive. See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a discussion of documented biases with regards to gender, religion, and race. Pythia-1B-deduped may produce socially unacceptable or undesirable text, *even if* the prompt itself does not include anything explicitly offensive. If you plan on using text generated through, for example, the Hosted Inference API, we recommend having a human curate the outputs of this language model before presenting it to other people. Please inform your audience that the text was generated by Pythia-1B-deduped. ### Quickstart Pythia models can be loaded and used via the following code, demonstrated here for the third `pythia-70m-deduped` checkpoint: ```python from transformers import GPTNeoXForCausalLM, AutoTokenizer model = GPTNeoXForCausalLM.from_pretrained( "EleutherAI/pythia-70m-deduped", revision="step3000", cache_dir="./pythia-70m-deduped/step3000", ) tokenizer = AutoTokenizer.from_pretrained( "EleutherAI/pythia-70m-deduped", revision="step3000", cache_dir="./pythia-70m-deduped/step3000", ) inputs = tokenizer("Hello, I am", return_tensors="pt") tokens = model.generate(**inputs) tokenizer.decode(tokens[0]) ``` Revision/branch `step143000` corresponds exactly to the model checkpoint on the `main` branch of each model.<br> For more information on how to use all Pythia models, see [documentation on GitHub](https://github.com/EleutherAI/pythia). ## Training ### Training data Pythia-1B-deduped was trained on the Pile **after the dataset has been globally deduplicated**.<br> [The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in English. It was created by EleutherAI specifically for training large language models. It contains texts from 22 diverse sources, roughly broken down into five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl), prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and miscellaneous (e.g. GitHub, Enron Emails). See [the Pile paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources, methodology, and a discussion of ethical implications. Consult [the datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation about the Pile and its component datasets. The Pile can be downloaded from the [official website](https://pile.eleuther.ai/), or from a [community mirror](https://the-eye.eu/public/AI/pile/). ### Training procedure All models were trained on the exact same data, in the exact same order. Each model saw 299,892,736,000 tokens during training, and 143 checkpoints for each model are saved every 2,097,152,000 tokens, spaced evenly throughout training, from `step1000` to `step143000` (which is the same as `main`). In addition, we also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`. This corresponds to training for just under 1 epoch on the Pile for non-deduplicated models, and about 1.5 epochs on the deduplicated Pile. All *Pythia* models trained for 143000 steps at a batch size of 2M (2,097,152 tokens).<br> See [GitHub](https://github.com/EleutherAI/pythia) for more details on training procedure, including [how to reproduce it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br> Pythia uses the same tokenizer as [GPT-NeoX- 20B](https://huggingface.co/EleutherAI/gpt-neox-20b). ## Evaluations All 16 *Pythia* models were evaluated using the [LM Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access the results by model and step at `results/json/*` in the [GitHub repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br> Expand the sections below to see plots of evaluation results for all Pythia and Pythia-deduped models compared with OPT and BLOOM. <details> <summary>LAMBADA – OpenAI</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/> </details> <details> <summary>Physical Interaction: Question Answering (PIQA)</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/> </details> <details> <summary>WinoGrande</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/> </details> <details> <summary>AI2 Reasoning Challenge—Easy Set</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/> </details> <details> <summary>SciQ</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/> </details> ## Changelog This section compares differences between previously released [Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current models. See Appendix B of the Pythia paper for further discussion of these changes and the motivation behind them. We found that retraining Pythia had no impact on benchmark performance. - All model sizes are now trained with uniform batch size of 2M tokens. Previously, the models of size 160M, 410M, and 1.4B parameters were trained with batch sizes of 4M tokens. - We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64, 128,256,512} in addition to every 1000 training steps. - Flash Attention was used in the new retrained suite. - We remedied a minor inconsistency that existed in the original suite: all models of size 2.8B parameters or smaller had a learning rate (LR) schedule which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and 12B models all used an LR schedule which decayed to a minimum LR of 0. In the redone training runs, we rectified this inconsistency: all models now were trained with LR decaying to a minimum of 0.1× their maximum LR. ### Naming convention and parameter count *Pythia* models were renamed in January 2023. It is possible that the old naming convention still persists in some documentation by accident. The current naming convention (70M, 160M, etc.) is based on total parameter count. <figure style="width:32em"> | current Pythia suffix | old suffix | total params | non-embedding params | | --------------------: | ---------: | -------------: | -------------------: | | 70M | 19M | 70,426,624 | 18,915,328 | | 160M | 125M | 162,322,944 | 85,056,000 | | 410M | 350M | 405,334,016 | 302,311,424 | | 1B | 800M | 1,011,781,632 | 805,736,448 | | 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 | | 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 | | 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 | | 12B | 13B | 11,846,072,320 | 11,327,027,200 | </figure>

sonoisa/sentence-bert-base-ja-mean-tokens

sonoisa

sentence-transformers

feature-extraction

--- language: ja license: cc-by-sa-4.0 tags: - sentence-transformers - sentence-bert - feature-extraction - sentence-similarity --- This is a Japanese sentence-BERT model. 日本語用Sentence-BERTモデル（バージョン1）です。 ※: 精度が1.5ポイントほど向上した[バージョン2モデル](https://huggingface.co/sonoisa/sentence-bert-base-ja-mean-tokens-v2)もあります。 # 解説 https://qiita.com/sonoisa/items/1df94d0a98cd4f209051 # 使い方 ```python from transformers import BertJapaneseTokenizer, BertModel import torch class SentenceBertJapanese: def __init__(self, model_name_or_path, device=None): self.tokenizer = BertJapaneseTokenizer.from_pretrained(model_name_or_path) self.model = BertModel.from_pretrained(model_name_or_path) self.model.eval() if device is None: device = "cuda" if torch.cuda.is_available() else "cpu" self.device = torch.device(device) self.model.to(device) def _mean_pooling(self, model_output, attention_mask): token_embeddings = model_output[0] #First element of model_output contains all token embeddings input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) @torch.no_grad() def encode(self, sentences, batch_size=8): all_embeddings = [] iterator = range(0, len(sentences), batch_size) for batch_idx in iterator: batch = sentences[batch_idx:batch_idx + batch_size] encoded_input = self.tokenizer.batch_encode_plus(batch, padding="longest", truncation=True, return_tensors="pt").to(self.device) model_output = self.model(**encoded_input) sentence_embeddings = self._mean_pooling(model_output, encoded_input["attention_mask"]).to('cpu') all_embeddings.extend(sentence_embeddings) # return torch.stack(all_embeddings).numpy() return torch.stack(all_embeddings) MODEL_NAME = "sonoisa/sentence-bert-base-ja-mean-tokens" model = SentenceBertJapanese(MODEL_NAME) sentences = ["暴走したAI", "暴走した人工知能"] sentence_embeddings = model.encode(sentences, batch_size=8) print("Sentence embeddings:", sentence_embeddings) ```

naver/efficient-splade-VI-BT-large-query

naver

transformers

fill-mask

--- license: cc-by-nc-sa-4.0 language: "en" tags: - splade - query-expansion - document-expansion - bag-of-words - passage-retrieval - knowledge-distillation - document encoder datasets: - ms_marco --- ## Efficient SPLADE Efficient SPLADE model for passage retrieval. This architecture uses two distinct models for query and document inference. This is the **query** one, please also download the **doc** one (https://huggingface.co/naver/efficient-splade-VI-BT-large-doc). For additional details, please visit: * paper: https://dl.acm.org/doi/10.1145/3477495.3531833 * code: https://github.com/naver/splade | | MRR@10 (MS MARCO dev) | R@1000 (MS MARCO dev) | Latency (PISA) ms | Latency (Inference) ms | --- | --- | --- | --- | --- | | `naver/efficient-splade-V-large` | 38.8 | 98.0 | 29.0 | 45.3 | `naver/efficient-splade-VI-BT-large` | 38.0 | 97.8 | 31.1 | 0.7 ## Citation If you use our checkpoint, please cite our work: ``` @inproceedings{10.1145/3477495.3531833, author = {Lassance, Carlos and Clinchant, St\'{e}phane}, title = {An Efficiency Study for SPLADE Models}, year = {2022}, isbn = {9781450387323}, publisher = {Association for Computing Machinery}, address = {New York, NY, USA}, url = {https://doi.org/10.1145/3477495.3531833}, doi = {10.1145/3477495.3531833}, abstract = {Latency and efficiency issues are often overlooked when evaluating IR models based on Pretrained Language Models (PLMs) in reason of multiple hardware and software testing scenarios. Nevertheless, efficiency is an important part of such systems and should not be overlooked. In this paper, we focus on improving the efficiency of the SPLADE model since it has achieved state-of-the-art zero-shot performance and competitive results on TREC collections. SPLADE efficiency can be controlled via a regularization factor, but solely controlling this regularization has been shown to not be efficient enough. In order to reduce the latency gap between SPLADE and traditional retrieval systems, we propose several techniques including L1 regularization for queries, a separation of document/query encoders, a FLOPS-regularized middle-training, and the use of faster query encoders. Our benchmark demonstrates that we can drastically improve the efficiency of these models while increasing the performance metrics on in-domain data. To our knowledge, we propose the first neural models that, under the same computing constraints, achieve similar latency (less than 4ms difference) as traditional BM25, while having similar performance (less than 10% MRR@10 reduction) as the state-of-the-art single-stage neural rankers on in-domain data.}, booktitle = {Proceedings of the 45th International ACM SIGIR Conference on Research and Development in Information Retrieval}, pages = {2220–2226}, numpages = {7}, keywords = {splade, latency, information retrieval, sparse representations}, location = {Madrid, Spain}, series = {SIGIR '22} } ```

timm/pnasnet5large.tf_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for pnasnet5large.tf_in1k A PNasNet image classification model. Trained on ImageNet-1k by paper authors. Ported from Tensorflow via Cadene's pretrained-models.pytorch. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 86.1 - GMACs: 25.0 - Activations (M): 92.9 - Image size: 331 x 331 - **Papers:** - Progressive Neural Architecture Search: https://arxiv.org/abs/1712.00559 - **Original:** - https://github.com/tensorflow/models - https://github.com/Cadene/pretrained-models.pytorch - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('pnasnet5large.tf_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'pnasnet5large.tf_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 96, 165, 165]) # torch.Size([1, 270, 83, 83]) # torch.Size([1, 1080, 42, 42]) # torch.Size([1, 2160, 21, 21]) # torch.Size([1, 4320, 11, 11]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'pnasnet5large.tf_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 4320, 11, 11) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @misc{liu2018progressive, title={Progressive Neural Architecture Search}, author={Chenxi Liu and Barret Zoph and Maxim Neumann and Jonathon Shlens and Wei Hua and Li-Jia Li and Li Fei-Fei and Alan Yuille and Jonathan Huang and Kevin Murphy}, year={2018}, eprint={1712.00559}, archivePrefix={arXiv}, primaryClass={cs.CV} } ```

timm/mobilevit_s.cvnets_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: other datasets: - imagenet-1k --- # Model card for mobilevit_s.cvnets_in1k A MobileViT image classification model. Trained on ImageNet-1k by paper authors. See license details at https://github.com/apple/ml-cvnets/blob/main/LICENSE ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 5.6 - GMACs: 2.0 - Activations (M): 19.9 - Image size: 256 x 256 - **Papers:** - MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer: https://arxiv.org/abs/2110.02178 - **Original:** https://github.com/apple/ml-cvnets - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('mobilevit_s.cvnets_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mobilevit_s.cvnets_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 32, 128, 128]) # torch.Size([1, 64, 64, 64]) # torch.Size([1, 96, 32, 32]) # torch.Size([1, 128, 16, 16]) # torch.Size([1, 640, 8, 8]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mobilevit_s.cvnets_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 640, 8, 8) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @inproceedings{mehta2022mobilevit, title={MobileViT: Light-weight, General-purpose, and Mobile-friendly Vision Transformer}, author={Sachin Mehta and Mohammad Rastegari}, booktitle={International Conference on Learning Representations}, year={2022} } ```

cross-encoder/nli-distilroberta-base

cross-encoder

transformers

zero-shot-classification

--- language: en pipeline_tag: zero-shot-classification tags: - distilroberta-base datasets: - multi_nli - snli metrics: - accuracy license: apache-2.0 --- # Cross-Encoder for Natural Language Inference This model was trained using [SentenceTransformers](https://sbert.net) [Cross-Encoder](https://www.sbert.net/examples/applications/cross-encoder/README.html) class. ## Training Data The model was trained on the [SNLI](https://nlp.stanford.edu/projects/snli/) and [MultiNLI](https://cims.nyu.edu/~sbowman/multinli/) datasets. For a given sentence pair, it will output three scores corresponding to the labels: contradiction, entailment, neutral. ## Performance For evaluation results, see [SBERT.net - Pretrained Cross-Encoder](https://www.sbert.net/docs/pretrained_cross-encoders.html#nli). ## Usage Pre-trained models can be used like this: ```python from sentence_transformers import CrossEncoder model = CrossEncoder('cross-encoder/nli-distilroberta-base') scores = model.predict([('A man is eating pizza', 'A man eats something'), ('A black race car starts up in front of a crowd of people.', 'A man is driving down a lonely road.')]) #Convert scores to labels label_mapping = ['contradiction', 'entailment', 'neutral'] labels = [label_mapping[score_max] for score_max in scores.argmax(axis=1)] ``` ## Usage with Transformers AutoModel You can use the model also directly with Transformers library (without SentenceTransformers library): ```python from transformers import AutoTokenizer, AutoModelForSequenceClassification import torch model = AutoModelForSequenceClassification.from_pretrained('cross-encoder/nli-distilroberta-base') tokenizer = AutoTokenizer.from_pretrained('cross-encoder/nli-distilroberta-base') features = tokenizer(['A man is eating pizza', 'A black race car starts up in front of a crowd of people.'], ['A man eats something', 'A man is driving down a lonely road.'], padding=True, truncation=True, return_tensors="pt") model.eval() with torch.no_grad(): scores = model(**features).logits label_mapping = ['contradiction', 'entailment', 'neutral'] labels = [label_mapping[score_max] for score_max in scores.argmax(dim=1)] print(labels) ``` ## Zero-Shot Classification This model can also be used for zero-shot-classification: ```python from transformers import pipeline classifier = pipeline("zero-shot-classification", model='cross-encoder/nli-distilroberta-base') sent = "Apple just announced the newest iPhone X" candidate_labels = ["technology", "sports", "politics"] res = classifier(sent, candidate_labels) print(res) ```

huggingface/informer-tourism-monthly

huggingface

transformers

--- license: apache-2.0 datasets: - monash_tsf --- # Informer ## Overview The Informer model was proposed in [Informer: Beyond Efficient Transformer for Long Sequence Time-Series Forecasting ](https://arxiv.org/abs/2012.07436) by Haoyi Zhou, Shanghang Zhang, Jieqi Peng, Shuai Zhang, Jianxin Li, Hui Xiong, and Wancai Zhang. This method introduces a Probabilistic Attention mechanism to select the "active" queries rather than the "lazy" queries and provides a sparse Transformer thus mitigating the quadratic compute and memory requirements of vanilla attention. The abstract from the paper is the following: *Many real-world applications require the prediction of long sequence time-series, such as electricity consumption planning. Long sequence time-series forecasting (LSTF) demands a high prediction capacity of the model, which is the ability to capture precise long-range dependency coupling between output and input efficiently. Recent studies have shown the potential of Transformer to increase the prediction capacity. However, there are several severe issues with Transformer that prevent it from being directly applicable to LSTF, including quadratic time complexity, high memory usage, and inherent limitation of the encoder-decoder architecture. To address these issues, we design an efficient transformer-based model for LSTF, named Informer, with three distinctive characteristics: (i) a ProbSparse self-attention mechanism, which achieves O(L logL) in time complexity and memory usage, and has comparable performance on sequences' dependency alignment. (ii) the self-attention distilling highlights dominating attention by halving cascading layer input, and efficiently handles extreme long input sequences. (iii) the generative style decoder, while conceptually simple, predicts the long time-series sequences at one forward operation rather than a step-by-step way, which drastically improves the inference speed of long-sequence predictions. Extensive experiments on four large-scale datasets demonstrate that Informer significantly outperforms existing methods and provides a new solution to the LSTF problem.*

timm/resnest101e.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for resnest101e.in1k A ResNeSt (ResNet based architecture with Split Attention) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 48.3 - GMACs: 13.4 - Activations (M): 28.7 - Image size: 256 x 256 - **Papers:** - ResNeSt: Split-Attention Networks: https://arxiv.org/abs/2004.08955 - **Dataset:** ImageNet-1k - **Original:** https://github.com/zhanghang1989/ResNeSt ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('resnest101e.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'resnest101e.in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 128, 128, 128]) # torch.Size([1, 256, 64, 64]) # torch.Size([1, 512, 32, 32]) # torch.Size([1, 1024, 16, 16]) # torch.Size([1, 2048, 8, 8]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'resnest101e.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 8, 8) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{zhang2020resnest, title={ResNeSt: Split-Attention Networks}, author={Zhang, Hang and Wu, Chongruo and Zhang, Zhongyue and Zhu, Yi and Zhang, Zhi and Lin, Haibin and Sun, Yue and He, Tong and Muller, Jonas and Manmatha, R. and Li, Mu and Smola, Alexander}, journal={arXiv preprint arXiv:2004.08955}, year={2020} } ```

sentence-transformers/msmarco-MiniLM-L-6-v3

sentence-transformers

sentence-transformers

sentence-similarity

--- license: apache-2.0 library_name: sentence-transformers tags: - sentence-transformers - feature-extraction - sentence-similarity - transformers pipeline_tag: sentence-similarity --- # sentence-transformers/msmarco-MiniLM-L-6-v3 This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 384 dimensional dense vector space and can be used for tasks like clustering or semantic search. ## Usage (Sentence-Transformers) Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed: ``` pip install -U sentence-transformers ``` Then you can use the model like this: ```python from sentence_transformers import SentenceTransformer sentences = ["This is an example sentence", "Each sentence is converted"] model = SentenceTransformer('sentence-transformers/msmarco-MiniLM-L-6-v3') embeddings = model.encode(sentences) print(embeddings) ``` ## Usage (HuggingFace Transformers) Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. ```python from transformers import AutoTokenizer, AutoModel import torch #Mean Pooling - Take attention mask into account for correct averaging def mean_pooling(model_output, attention_mask): token_embeddings = model_output[0] #First element of model_output contains all token embeddings input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) # Sentences we want sentence embeddings for sentences = ['This is an example sentence', 'Each sentence is converted'] # Load model from HuggingFace Hub tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/msmarco-MiniLM-L-6-v3') model = AutoModel.from_pretrained('sentence-transformers/msmarco-MiniLM-L-6-v3') # Tokenize sentences encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt') # Compute token embeddings with torch.no_grad(): model_output = model(**encoded_input) # Perform pooling. In this case, max pooling. sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask']) print("Sentence embeddings:") print(sentence_embeddings) ``` ## Evaluation Results For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/msmarco-MiniLM-L-6-v3) ## Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: BertModel (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False}) ) ``` ## Citing & Authors This model was trained by [sentence-transformers](https://www.sbert.net/). If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084): ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "http://arxiv.org/abs/1908.10084", } ```

sentence-transformers/msmarco-distilbert-dot-v5

sentence-transformers

sentence-transformers

sentence-similarity

--- language: - en license: apache-2.0 library_name: sentence-transformers tags: - sentence-transformers - feature-extraction - sentence-similarity - transformers pipeline_tag: sentence-similarity --- # msmarco-distilbert-dot-v5 This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and was designed for **semantic search**. It has been trained on 500K (query, answer) pairs from the [MS MARCO dataset](https://github.com/microsoft/MSMARCO-Passage-Ranking/). For an introduction to semantic search, have a look at: [SBERT.net - Semantic Search](https://www.sbert.net/examples/applications/semantic-search/README.html) ## Usage (Sentence-Transformers) Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed: ``` pip install -U sentence-transformers ``` Then you can use the model like this: ```python from sentence_transformers import SentenceTransformer, util query = "How many people live in London?" docs = ["Around 9 Million people live in London", "London is known for its financial district"] #Load the model model = SentenceTransformer('sentence-transformers/msmarco-distilbert-dot-v5') #Encode query and documents query_emb = model.encode(query) doc_emb = model.encode(docs) #Compute dot score between query and all document embeddings scores = util.dot_score(query_emb, doc_emb)[0].cpu().tolist() #Combine docs & scores doc_score_pairs = list(zip(docs, scores)) #Sort by decreasing score doc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True) #Output passages & scores print("Query:", query) for doc, score in doc_score_pairs: print(score, doc) ``` ## Usage (HuggingFace Transformers) Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the correct pooling-operation on-top of the contextualized word embeddings. ```python from transformers import AutoTokenizer, AutoModel import torch #Mean Pooling - Take attention mask into account for correct averaging def mean_pooling(model_output, attention_mask): token_embeddings = model_output.last_hidden_state input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) #Encode text def encode(texts): # Tokenize sentences encoded_input = tokenizer(texts, padding=True, truncation=True, return_tensors='pt') # Compute token embeddings with torch.no_grad(): model_output = model(**encoded_input, return_dict=True) # Perform pooling embeddings = mean_pooling(model_output, encoded_input['attention_mask']) return embeddings # Sentences we want sentence embeddings for query = "How many people live in London?" docs = ["Around 9 Million people live in London", "London is known for its financial district"] # Load model from HuggingFace Hub tokenizer = AutoTokenizer.from_pretrained("sentence-transformers/msmarco-distilbert-dot-v5") model = AutoModel.from_pretrained("sentence-transformers/msmarco-distilbert-dot-v5") #Encode query and docs query_emb = encode(query) doc_emb = encode(docs) #Compute dot score between query and all document embeddings scores = torch.mm(query_emb, doc_emb.transpose(0, 1))[0].cpu().tolist() #Combine docs & scores doc_score_pairs = list(zip(docs, scores)) #Sort by decreasing score doc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True) #Output passages & scores print("Query:", query) for doc, score in doc_score_pairs: print(score, doc) ``` ## Technical Details In the following some technical details how this model must be used: | Setting | Value | | --- | :---: | | Dimensions | 768 | | Max Sequence Length | 512 | | Produces normalized embeddings | No | | Pooling-Method | Mean pooling | | Suitable score functions | dot-product (e.g. `util.dot_score`) | ## Evaluation Results <!--- Describe how your model was evaluated --> For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=msmarco-distilbert-base-dot-v5) ## Training See `train_script.py` in this repository for the used training script. The model was trained with the parameters: **DataLoader**: `torch.utils.data.dataloader.DataLoader` of length 7858 with parameters: ``` {'batch_size': 64, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'} ``` **Loss**: `sentence_transformers.losses.MarginMSELoss.MarginMSELoss` Parameters of the fit()-Method: ``` { "callback": null, "epochs": 30, "evaluation_steps": 0, "evaluator": "NoneType", "max_grad_norm": 1, "optimizer_class": "<class 'transformers.optimization.AdamW'>", "optimizer_params": { "lr": 1e-05 }, "scheduler": "WarmupLinear", "steps_per_epoch": null, "warmup_steps": 10000, "weight_decay": 0.01 } ``` ## Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: DistilBertModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False}) ) ``` ## Citing & Authors This model was trained by [sentence-transformers](https://www.sbert.net/). If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084): ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "http://arxiv.org/abs/1908.10084", } ``` ## License This model is released under the Apache 2 license. However, note that this model was trained on the MS MARCO dataset which has it's own license restrictions: [MS MARCO - Terms and Conditions](https://github.com/microsoft/msmarco/blob/095515e8e28b756a62fcca7fcf1d8b3d9fbb96a9/README.md).

EleutherAI/pythia-1.4b

EleutherAI

transformers

text-generation

--- language: - en tags: - pytorch - causal-lm - pythia license: apache-2.0 datasets: - EleutherAI/the_pile --- The *Pythia Scaling Suite* is a collection of models developed to facilitate interpretability research [(see paper)](https://arxiv.org/pdf/2304.01373.pdf). It contains two sets of eight models of sizes 70M, 160M, 410M, 1B, 1.4B, 2.8B, 6.9B, and 12B. For each size, there are two models: one trained on the Pile, and one trained on the Pile after the dataset has been globally deduplicated. All 8 model sizes are trained on the exact same data, in the exact same order. We also provide 154 intermediate checkpoints per model, hosted on Hugging Face as branches. The Pythia model suite was deliberately designed to promote scientific research on large language models, especially interpretability research. Despite not centering downstream performance as a design goal, we find the models <a href="#evaluations">match or exceed</a> the performance of similar and same-sized models, such as those in the OPT and GPT-Neo suites. <details> <summary style="font-weight:600">Details on previous early release and naming convention.</summary> Previously, we released an early version of the Pythia suite to the public. However, we decided to retrain the model suite to address a few hyperparameter discrepancies. This model card <a href="#changelog">lists the changes</a>; see appendix B in the Pythia paper for further discussion. We found no difference in benchmark performance between the two Pythia versions. The old models are [still available](https://huggingface.co/models?other=pythia_v0), but we suggest the retrained suite if you are just starting to use Pythia.<br> **This is the current release.** Please note that all models in the *Pythia* suite were renamed in January 2023. For clarity, a <a href="#naming-convention-and-parameter-count">table comparing the old and new names</a> is provided in this model card, together with exact parameter counts. </details> <br> # Pythia-1.4B ## Model Details - Developed by: [EleutherAI](http://eleuther.ai) - Model type: Transformer-based Language Model - Language: English - Learn more: [Pythia's GitHub repository](https://github.com/EleutherAI/pythia) for training procedure, config files, and details on how to use. [See paper](https://arxiv.org/pdf/2304.01373.pdf) for more evals and implementation details. - Library: [GPT-NeoX](https://github.com/EleutherAI/gpt-neox) - License: Apache 2.0 - Contact: to ask questions about this model, join the [EleutherAI Discord](https://discord.gg/zBGx3azzUn), and post them in `#release-discussion`. Please read the existing *Pythia* documentation before asking about it in the EleutherAI Discord. For general correspondence: [contact@eleuther. ai](mailto:contact@eleuther.ai). <figure> | Pythia model | Non-Embedding Params | Layers | Model Dim | Heads | Batch Size | Learning Rate | Equivalent Models | | -----------: | -------------------: | :----: | :-------: | :---: | :--------: | :-------------------: | :--------------------: | | 70M | 18,915,328 | 6 | 512 | 8 | 2M | 1.0 x 10^-3 | — | | 160M | 85,056,000 | 12 | 768 | 12 | 2M | 6.0 x 10^-4 | GPT-Neo 125M, OPT-125M | | 410M | 302,311,424 | 24 | 1024 | 16 | 2M | 3.0 x 10^-4 | OPT-350M | | 1.0B | 805,736,448 | 16 | 2048 | 8 | 2M | 3.0 x 10^-4 | — | | 1.4B | 1,208,602,624 | 24 | 2048 | 16 | 2M | 2.0 x 10^-4 | GPT-Neo 1.3B, OPT-1.3B | | 2.8B | 2,517,652,480 | 32 | 2560 | 32 | 2M | 1.6 x 10^-4 | GPT-Neo 2.7B, OPT-2.7B | | 6.9B | 6,444,163,072 | 32 | 4096 | 32 | 2M | 1.2 x 10^-4 | OPT-6.7B | | 12B | 11,327,027,200 | 36 | 5120 | 40 | 2M | 1.2 x 10^-4 | — | <figcaption>Engineering details for the <i>Pythia Suite</i>. Deduped and non-deduped models of a given size have the same hyperparameters. “Equivalent” models have <b>exactly</b> the same architecture, and the same number of non-embedding parameters.</figcaption> </figure> ## Uses and Limitations ### Intended Use The primary intended use of Pythia is research on the behavior, functionality, and limitations of large language models. This suite is intended to provide a controlled setting for performing scientific experiments. We also provide 154 checkpoints per model: initial `step0`, 10 log-spaced checkpoints `step{1,2,4...512}`, and 143 evenly-spaced checkpoints from `step1000` to `step143000`. These checkpoints are hosted on Hugging Face as branches. Note that branch `143000` corresponds exactly to the model checkpoint on the `main` branch of each model. You may also further fine-tune and adapt Pythia-1.4B for deployment, as long as your use is in accordance with the Apache 2.0 license. Pythia models work with the Hugging Face [Transformers Library](https://huggingface.co/docs/transformers/index). If you decide to use pre-trained Pythia-1.4B as a basis for your fine-tuned model, please conduct your own risk and bias assessment. ### Out-of-scope use The Pythia Suite is **not** intended for deployment. It is not a in itself a product and cannot be used for human-facing interactions. For example, the model may generate harmful or offensive text. Please evaluate the risks associated with your particular use case. Pythia models are English-language only, and are not suitable for translation or generating text in other languages. Pythia-1.4B has not been fine-tuned for downstream contexts in which language models are commonly deployed, such as writing genre prose, or commercial chatbots. This means Pythia-1.4B will **not** respond to a given prompt the way a product like ChatGPT does. This is because, unlike this model, ChatGPT was fine-tuned using methods such as Reinforcement Learning from Human Feedback (RLHF) to better “follow” human instructions. ### Limitations and biases The core functionality of a large language model is to take a string of text and predict the next token. The token used by the model need not produce the most “accurate” text. Never rely on Pythia-1.4B to produce factually accurate output. This model was trained on [the Pile](https://pile.eleuther.ai/), a dataset known to contain profanity and texts that are lewd or otherwise offensive. See [Section 6 of the Pile paper](https://arxiv.org/abs/2101.00027) for a discussion of documented biases with regards to gender, religion, and race. Pythia-1.4B may produce socially unacceptable or undesirable text, *even if* the prompt itself does not include anything explicitly offensive. If you plan on using text generated through, for example, the Hosted Inference API, we recommend having a human curate the outputs of this language model before presenting it to other people. Please inform your audience that the text was generated by Pythia-1.4B. ### Quickstart Pythia models can be loaded and used via the following code, demonstrated here for the third `pythia-70m-deduped` checkpoint: ```python from transformers import GPTNeoXForCausalLM, AutoTokenizer model = GPTNeoXForCausalLM.from_pretrained( "EleutherAI/pythia-70m-deduped", revision="step3000", cache_dir="./pythia-70m-deduped/step3000", ) tokenizer = AutoTokenizer.from_pretrained( "EleutherAI/pythia-70m-deduped", revision="step3000", cache_dir="./pythia-70m-deduped/step3000", ) inputs = tokenizer("Hello, I am", return_tensors="pt") tokens = model.generate(**inputs) tokenizer.decode(tokens[0]) ``` Revision/branch `step143000` corresponds exactly to the model checkpoint on the `main` branch of each model.<br> For more information on how to use all Pythia models, see [documentation on GitHub](https://github.com/EleutherAI/pythia). ## Training ### Training data [The Pile](https://pile.eleuther.ai/) is a 825GiB general-purpose dataset in English. It was created by EleutherAI specifically for training large language models. It contains texts from 22 diverse sources, roughly broken down into five categories: academic writing (e.g. arXiv), internet (e.g. CommonCrawl), prose (e.g. Project Gutenberg), dialogue (e.g. YouTube subtitles), and miscellaneous (e.g. GitHub, Enron Emails). See [the Pile paper](https://arxiv.org/abs/2101.00027) for a breakdown of all data sources, methodology, and a discussion of ethical implications. Consult [the datasheet](https://arxiv.org/abs/2201.07311) for more detailed documentation about the Pile and its component datasets. The Pile can be downloaded from the [official website](https://pile.eleuther.ai/), or from a [community mirror](https://the-eye.eu/public/AI/pile/).<br> The Pile was **not** deduplicated before being used to train Pythia-1.4B. ### Training procedure All models were trained on the exact same data, in the exact same order. Each model saw 299,892,736,000 tokens during training, and 143 checkpoints for each model are saved every 2,097,152,000 tokens, spaced evenly throughout training, from `step1000` to `step143000` (which is the same as `main`). In addition, we also provide frequent early checkpoints: `step0` and `step{1,2,4...512}`. This corresponds to training for just under 1 epoch on the Pile for non-deduplicated models, and about 1.5 epochs on the deduplicated Pile. All *Pythia* models trained for 143000 steps at a batch size of 2M (2,097,152 tokens).<br> See [GitHub](https://github.com/EleutherAI/pythia) for more details on training procedure, including [how to reproduce it](https://github.com/EleutherAI/pythia/blob/main/README.md#reproducing-training).<br> Pythia uses the same tokenizer as [GPT-NeoX- 20B](https://huggingface.co/EleutherAI/gpt-neox-20b). ## Evaluations All 16 *Pythia* models were evaluated using the [LM Evaluation Harness](https://github.com/EleutherAI/lm-evaluation-harness). You can access the results by model and step at `results/json/*` in the [GitHub repository](https://github.com/EleutherAI/pythia/tree/main/results/json/).<br> Expand the sections below to see plots of evaluation results for all Pythia and Pythia-deduped models compared with OPT and BLOOM. <details> <summary>LAMBADA – OpenAI</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/lambada_openai_v1.png" style="width:auto"/> </details> <details> <summary>Physical Interaction: Question Answering (PIQA)</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/piqa_v1.png" style="width:auto"/> </details> <details> <summary>WinoGrande</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/winogrande_v1.png" style="width:auto"/> </details> <details> <summary>AI2 Reasoning Challenge—Easy Set</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/arc_easy_v1.png" style="width:auto"/> </details> <details> <summary>SciQ</summary> <img src="/EleutherAI/pythia-12b/resolve/main/eval_plots/sciq_v1.png" style="width:auto"/> </details> ## Changelog This section compares differences between previously released [Pythia v0](https://huggingface.co/models?other=pythia_v0) and the current models. See Appendix B of the Pythia paper for further discussion of these changes and the motivation behind them. We found that retraining Pythia had no impact on benchmark performance. - All model sizes are now trained with uniform batch size of 2M tokens. Previously, the models of size 160M, 410M, and 1.4B parameters were trained with batch sizes of 4M tokens. - We added checkpoints at initialization (step 0) and steps {1,2,4,8,16,32,64, 128,256,512} in addition to every 1000 training steps. - Flash Attention was used in the new retrained suite. - We remedied a minor inconsistency that existed in the original suite: all models of size 2.8B parameters or smaller had a learning rate (LR) schedule which decayed to a minimum LR of 10% the starting LR rate, but the 6.9B and 12B models all used an LR schedule which decayed to a minimum LR of 0. In the redone training runs, we rectified this inconsistency: all models now were trained with LR decaying to a minimum of 0.1× their maximum LR. ### Naming convention and parameter count *Pythia* models were renamed in January 2023. It is possible that the old naming convention still persists in some documentation by accident. The current naming convention (70M, 160M, etc.) is based on total parameter count. <figure style="width:32em"> | current Pythia suffix | old suffix | total params | non-embedding params | | --------------------: | ---------: | -------------: | -------------------: | | 70M | 19M | 70,426,624 | 18,915,328 | | 160M | 125M | 162,322,944 | 85,056,000 | | 410M | 350M | 405,334,016 | 302,311,424 | | 1B | 800M | 1,011,781,632 | 805,736,448 | | 1.4B | 1.3B | 1,414,647,808 | 1,208,602,624 | | 2.8B | 2.7B | 2,775,208,960 | 2,517,652,480 | | 6.9B | 6.7B | 6,857,302,016 | 6,444,163,072 | | 12B | 13B | 11,846,072,320 | 11,327,027,200 | </figure>

owkin/phikon

owkin

transformers

feature-extraction

--- license: other language: - en tags: - biology - medical - cancer datasets: - owkin/nct-crc-he - owkin/camelyon16-features pipeline_tag: feature-extraction --- # Model Card for Phikon --- Phikon is a self-supervised learning model for histopathology trained with iBOT. To learn more about how to use the model, we encourage you to read our blog post and view this Colab notebook. ### Model Description - **Developed by:** Owkin - **Funded by:** Owkin and IDRIS - **Model type:** Vision Transformer Base - **Model Stats:** - Params (M): 85.8 - Image size: 224 x 224 x 3 - **Paper:** - Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling. A. Filiot et al., medRxiv 2023.07.21.23292757; doi: [https://doi.org/10.1101/2023.07.21.23292757](https://www.medrxiv.org/content/10.1101/2023.07.21.23292757v2) - **Pretrain Dataset:** 40 million pan-cancer tiles extracted from [TGCA](https://portal.gdc.cancer.gov/) - **Original:** https://github.com/owkin/HistoSSLscaling/ - **License:** [Owkin non-commercial license](https://github.com/owkin/HistoSSLscaling/blob/main/LICENSE.txt) ## Uses ### Direct Use The primary use of the Phikon model can be used for feature extraction from histology image tiles. ### Downstream Use The model can be used for cancer classification on a variety of cancer subtypes. The model can also be finetuned to specialise on cancer subtypes. ## Technical Specifications ### Compute Infrastructure All the models we built were trained on the French Jean Zay cluster. ### Hardware NVIDIA V100 GPUs with 32Gb RAM ### Software PyTorch 1.13.1 --- ### BibTeX entry and citation info ```bibtex @article{Filiot2023ScalingSSLforHistoWithMIM, author = {Alexandre Filiot and Ridouane Ghermi and Antoine Olivier and Paul Jacob and Lucas Fidon and Alice Mac Kain and Charlie Saillard and Jean-Baptiste Schiratti}, title = {Scaling Self-Supervised Learning for Histopathology with Masked Image Modeling}, elocation-id = {2023.07.21.23292757}, year = {2023}, doi = {10.1101/2023.07.21.23292757}, publisher = {Cold Spring Harbor Laboratory Press}, url = {https://www.medrxiv.org/content/early/2023/07/26/2023.07.21.23292757}, eprint = {https://www.medrxiv.org/content/early/2023/07/26/2023.07.21.23292757.full.pdf}, journal = {medRxiv} } ```

katuni4ka/tiny-random-baichuan2

katuni4ka

transformers

text-generation

Entry not found

google-bert/bert-base-german-dbmdz-uncased

google-bert

transformers

fill-mask

--- language: de license: mit --- This model is the same as [dbmdz/bert-base-german-uncased](https://huggingface.co/dbmdz/bert-base-german-uncased). See the [dbmdz/bert-base-german-cased model card](https://huggingface.co/dbmdz/bert-base-german-uncased) for details on the model.

timm/spnasnet_100.rmsp_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for spnasnet_100.rmsp_in1k A SPNasNet image classification model. Trained on ImageNet-1k in `timm` using recipe template described below. Recipe details: * A simple RmsProp based recipe without RandAugment. Using RandomErasing, mixup, dropout, standard random-resize-crop augmentation. * RMSProp (TF 1.0 behaviour) optimizer, EMA weight averaging * Step (exponential decay w/ staircase) LR schedule with warmup ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 4.4 - GMACs: 0.3 - Activations (M): 6.0 - Image size: 224 x 224 - **Papers:** - Single-Path NAS: Designing Hardware-Efficient ConvNets in less than 4 Hours: https://arxiv.org/abs/1904.02877 - **Dataset:** ImageNet-1k - **Original:** https://github.com/huggingface/pytorch-image-models ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('spnasnet_100.rmsp_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'spnasnet_100.rmsp_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 16, 112, 112]) # torch.Size([1, 24, 56, 56]) # torch.Size([1, 40, 28, 28]) # torch.Size([1, 96, 14, 14]) # torch.Size([1, 320, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'spnasnet_100.rmsp_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1280, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ``` ```bibtex @inproceedings{stamoulis2020single, title={Single-path nas: Designing hardware-efficient convnets in less than 4 hours}, author={Stamoulis, Dimitrios and Ding, Ruizhou and Wang, Di and Lymberopoulos, Dimitrios and Priyantha, Bodhi and Liu, Jie and Marculescu, Diana}, booktitle={Machine Learning and Knowledge Discovery in Databases: European Conference, ECML PKDD 2019, W{"u}rzburg, Germany, September 16--20, 2019, Proceedings, Part II}, pages={481--497}, year={2020}, organization={Springer} } ```

suno/bark-small

suno

transformers

text-to-speech

--- language: - en - de - es - fr - hi - it - ja - ko - pl - pt - ru - tr - zh thumbnail: >- https://user-images.githubusercontent.com/5068315/230698495-cbb1ced9-c911-4c9a-941d-a1a4a1286ac6.png library: bark license: mit tags: - bark - audio - text-to-speech duplicated_from: ylacombe/bark-small pipeline_tag: text-to-speech --- # Bark Bark is a transformer-based text-to-audio model created by [Suno](https://www.suno.ai). Bark can generate highly realistic, multilingual speech as well as other audio - including music, background noise and simple sound effects. The model can also produce nonverbal communications like laughing, sighing and crying. To support the research community, we are providing access to pretrained model checkpoints ready for inference. The original github repo and model card can be found [here](https://github.com/suno-ai/bark). This model is meant for research purposes only. The model output is not censored and the authors do not endorse the opinions in the generated content. Use at your own risk. Two checkpoints are released: - [**small** (this checkpoint)](https://huggingface.co/suno/bark-small) - [large](https://huggingface.co/suno/bark) ## Example Try out Bark yourself! * Bark Colab: <a target="_blank" href="https://colab.research.google.com/drive/1eJfA2XUa-mXwdMy7DoYKVYHI1iTd9Vkt?usp=sharing"> <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/> </a> * Hugging Face Colab: <a target="_blank" href="https://colab.research.google.com/drive/1dWWkZzvu7L9Bunq9zvD-W02RFUXoW-Pd?usp=sharing"> <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/> </a> * Hugging Face Demo: <a target="_blank" href="https://huggingface.co/spaces/suno/bark"> <img src="https://huggingface.co/datasets/huggingface/badges/raw/main/open-in-hf-spaces-sm.svg" alt="Open in HuggingFace"/> </a> ## 🤗 Transformers Usage You can run Bark locally with the 🤗 Transformers library from version 4.31.0 onwards. 1. First install the 🤗 [Transformers library](https://github.com/huggingface/transformers) and scipy: ``` pip install --upgrade pip pip install --upgrade transformers scipy ``` 2. Run inference via the `Text-to-Speech` (TTS) pipeline. You can infer the bark model via the TTS pipeline in just a few lines of code! ```python from transformers import pipeline import scipy synthesiser = pipeline("text-to-speech", "suno/bark-small") speech = synthesiser("Hello, my dog is cooler than you!", forward_params={"do_sample": True}) scipy.io.wavfile.write("bark_out.wav", rate=speech["sampling_rate"], data=speech["audio"]) ``` 3. Run inference via the Transformers modelling code. You can use the processor + generate code to convert text into a mono 24 kHz speech waveform for more fine-grained control. ```python from transformers import AutoProcessor, AutoModel processor = AutoProcessor.from_pretrained("suno/bark-small") model = AutoModel.from_pretrained("suno/bark-small") inputs = processor( text=["Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe."], return_tensors="pt", ) speech_values = model.generate(**inputs, do_sample=True) ``` 4. Listen to the speech samples either in an ipynb notebook: ```python from IPython.display import Audio sampling_rate = model.generation_config.sample_rate Audio(speech_values.cpu().numpy().squeeze(), rate=sampling_rate) ``` Or save them as a `.wav` file using a third-party library, e.g. `scipy`: ```python import scipy sampling_rate = model.config.sample_rate scipy.io.wavfile.write("bark_out.wav", rate=sampling_rate, data=speech_values.cpu().numpy().squeeze()) ``` For more details on using the Bark model for inference using the 🤗 Transformers library, refer to the [Bark docs](https://huggingface.co/docs/transformers/model_doc/bark). ### Optimization tips Refers to this [blog post](https://huggingface.co/blog/optimizing-bark#benchmark-results) to find out more about the following methods and a benchmark of their benefits. #### Get significant speed-ups: **Using 🤗 Better Transformer** Better Transformer is an 🤗 Optimum feature that performs kernel fusion under the hood. You can gain 20% to 30% in speed with zero performance degradation. It only requires one line of code to export the model to 🤗 Better Transformer: ```python model = model.to_bettertransformer() ``` Note that 🤗 Optimum must be installed before using this feature. [Here's how to install it.](https://huggingface.co/docs/optimum/installation) **Using Flash Attention 2** Flash Attention 2 is an even faster, optimized version of the previous optimization. ```python model = BarkModel.from_pretrained("suno/bark-small", torch_dtype=torch.float16, use_flash_attention_2=True).to(device) ``` Make sure to load your model in half-precision (e.g. `torch.float16``) and to [install](https://github.com/Dao-AILab/flash-attention#installation-and-features) the latest version of Flash Attention 2. **Note:** Flash Attention 2 is only available on newer GPUs, refer to 🤗 Better Transformer in case your GPU don't support it. #### Reduce memory footprint: **Using half-precision** You can speed up inference and reduce memory footprint by 50% simply by loading the model in half-precision (e.g. `torch.float16``). **Using CPU offload** Bark is made up of 4 sub-models, which are called up sequentially during audio generation. In other words, while one sub-model is in use, the other sub-models are idle. If you're using a CUDA device, a simple solution to benefit from an 80% reduction in memory footprint is to offload the GPU's submodels when they're idle. This operation is called CPU offloading. You can use it with one line of code. ```python model.enable_cpu_offload() ``` Note that 🤗 Accelerate must be installed before using this feature. [Here's how to install it.](https://huggingface.co/docs/accelerate/basic_tutorials/install) ## Suno Usage You can also run Bark locally through the original [Bark library]((https://github.com/suno-ai/bark): 1. First install the [`bark` library](https://github.com/suno-ai/bark) 3. Run the following Python code: ```python from bark import SAMPLE_RATE, generate_audio, preload_models from IPython.display import Audio # download and load all models preload_models() # generate audio from text text_prompt = """ Hello, my name is Suno. And, uh — and I like pizza. [laughs] But I also have other interests such as playing tic tac toe. """ speech_array = generate_audio(text_prompt) # play text in notebook Audio(speech_array, rate=SAMPLE_RATE) ``` [pizza.webm](https://user-images.githubusercontent.com/5068315/230490503-417e688d-5115-4eee-9550-b46a2b465ee3.webm) To save `audio_array` as a WAV file: ```python from scipy.io.wavfile import write as write_wav write_wav("/path/to/audio.wav", SAMPLE_RATE, audio_array) ``` ## Model Details The following is additional information about the models released here. Bark is a series of three transformer models that turn text into audio. ### Text to semantic tokens - Input: text, tokenized with [BERT tokenizer from Hugging Face](https://huggingface.co/docs/transformers/model_doc/bert#transformers.BertTokenizer) - Output: semantic tokens that encode the audio to be generated ### Semantic to coarse tokens - Input: semantic tokens - Output: tokens from the first two codebooks of the [EnCodec Codec](https://github.com/facebookresearch/encodec) from facebook ### Coarse to fine tokens - Input: the first two codebooks from EnCodec - Output: 8 codebooks from EnCodec ### Architecture | Model | Parameters | Attention | Output Vocab size | |:-------------------------:|:----------:|------------|:-----------------:| | Text to semantic tokens | 80/300 M | Causal | 10,000 | | Semantic to coarse tokens | 80/300 M | Causal | 2x 1,024 | | Coarse to fine tokens | 80/300 M | Non-causal | 6x 1,024 | ### Release date April 2023 ## Broader Implications We anticipate that this model's text to audio capabilities can be used to improve accessbility tools in a variety of languages. While we hope that this release will enable users to express their creativity and build applications that are a force for good, we acknowledge that any text to audio model has the potential for dual use. While it is not straightforward to voice clone known people with Bark, it can still be used for nefarious purposes. To further reduce the chances of unintended use of Bark, we also release a simple classifier to detect Bark-generated audio with high accuracy (see notebooks section of the main repository). ## License Bark is licensed under the [MIT License](https://github.com/suno-ai/bark/blob/main/LICENSE), meaning it's available for commercial use.

Qwen/Qwen1.5-7B

Qwen

transformers

text-generation

--- license: other license_name: tongyi-qianwen license_link: >- https://huggingface.co/Qwen/Qwen1.5-7B/blob/main/LICENSE language: - en pipeline_tag: text-generation tags: - pretrained --- # Qwen1.5-7B ## Introduction Qwen1.5 is the beta version of Qwen2, a transformer-based decoder-only language model pretrained on a large amount of data. In comparison with the previous released Qwen, the improvements include: * 8 model sizes, including 0.5B, 1.8B, 4B, 7B, 14B, 32B and 72B dense models, and an MoE model of 14B with 2.7B activated; * Significant performance improvement in Chat models; * Multilingual support of both base and chat models; * Stable support of 32K context length for models of all sizes * No need of `trust_remote_code`. For more details, please refer to our [blog post](https://qwenlm.github.io/blog/qwen1.5/) and [GitHub repo](https://github.com/QwenLM/Qwen1.5). ## Model Details Qwen1.5 is a language model series including decoder language models of different model sizes. For each size, we release the base language model and the aligned chat model. It is based on the Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, mixture of sliding window attention and full attention, etc. Additionally, we have an improved tokenizer adaptive to multiple natural languages and codes. For the beta version, temporarily we did not include GQA (except for 32B) and the mixture of SWA and full attention. ## Requirements The code of Qwen1.5 has been in the latest Hugging face transformers and we advise you to install `transformers>=4.37.0`, or you might encounter the following error: ``` KeyError: 'qwen2'. ``` ## Usage We do not advise you to use base language models for text generation. Instead, you can apply post-training, e.g., SFT, RLHF, continued pretraining, etc., on this model. ## Citation If you find our work helpful, feel free to give us a cite. ``` @article{qwen, title={Qwen Technical Report}, author={Jinze Bai and Shuai Bai and Yunfei Chu and Zeyu Cui and Kai Dang and Xiaodong Deng and Yang Fan and Wenbin Ge and Yu Han and Fei Huang and Binyuan Hui and Luo Ji and Mei Li and Junyang Lin and Runji Lin and Dayiheng Liu and Gao Liu and Chengqiang Lu and Keming Lu and Jianxin Ma and Rui Men and Xingzhang Ren and Xuancheng Ren and Chuanqi Tan and Sinan Tan and Jianhong Tu and Peng Wang and Shijie Wang and Wei Wang and Shengguang Wu and Benfeng Xu and Jin Xu and An Yang and Hao Yang and Jian Yang and Shusheng Yang and Yang Yao and Bowen Yu and Hongyi Yuan and Zheng Yuan and Jianwei Zhang and Xingxuan Zhang and Yichang Zhang and Zhenru Zhang and Chang Zhou and Jingren Zhou and Xiaohuan Zhou and Tianhang Zhu}, journal={arXiv preprint arXiv:2309.16609}, year={2023} } ```

dmis-lab/biobert-base-cased-v1.2

dmis-lab

transformers

fill-mask

Entry not found

deepseek-ai/deepseek-coder-6.7b-instruct

deepseek-ai

transformers

text-generation

--- license: other license_name: deepseek license_link: LICENSE --- <p align="center"> <img width="1000px" alt="DeepSeek Coder" src="https://github.com/deepseek-ai/DeepSeek-Coder/blob/main/pictures/logo.png?raw=true"> </p> <p align="center"><a href="https://www.deepseek.com/">[🏠Homepage]</a> | <a href="https://coder.deepseek.com/">[🤖 Chat with DeepSeek Coder]</a> | <a href="https://discord.gg/Tc7c45Zzu5">[Discord]</a> | <a href="https://github.com/guoday/assert/blob/main/QR.png?raw=true">[Wechat(微信)]</a> </p> <hr> ### 1. Introduction of Deepseek Coder Deepseek Coder is composed of a series of code language models, each trained from scratch on 2T tokens, with a composition of 87% code and 13% natural language in both English and Chinese. We provide various sizes of the code model, ranging from 1B to 33B versions. Each model is pre-trained on project-level code corpus by employing a window size of 16K and a extra fill-in-the-blank task, to support project-level code completion and infilling. For coding capabilities, Deepseek Coder achieves state-of-the-art performance among open-source code models on multiple programming languages and various benchmarks. - **Massive Training Data**: Trained from scratch fon 2T tokens, including 87% code and 13% linguistic data in both English and Chinese languages. - **Highly Flexible & Scalable**: Offered in model sizes of 1.3B, 5.7B, 6.7B, and 33B, enabling users to choose the setup most suitable for their requirements. - **Superior Model Performance**: State-of-the-art performance among publicly available code models on HumanEval, MultiPL-E, MBPP, DS-1000, and APPS benchmarks. - **Advanced Code Completion Capabilities**: A window size of 16K and a fill-in-the-blank task, supporting project-level code completion and infilling tasks. ### 2. Model Summary deepseek-coder-6.7b-instruct is a 6.7B parameter model initialized from deepseek-coder-6.7b-base and fine-tuned on 2B tokens of instruction data. - **Home Page:** [DeepSeek](https://deepseek.com/) - **Repository:** [deepseek-ai/deepseek-coder](https://github.com/deepseek-ai/deepseek-coder) - **Chat With DeepSeek Coder:** [DeepSeek-Coder](https://coder.deepseek.com/) ### 3. How to Use Here give some examples of how to use our model. #### Chat Model Inference ```python from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("deepseek-ai/deepseek-coder-6.7b-instruct", trust_remote_code=True) model = AutoModelForCausalLM.from_pretrained("deepseek-ai/deepseek-coder-6.7b-instruct", trust_remote_code=True, torch_dtype=torch.bfloat16).cuda() messages=[ { 'role': 'user', 'content': "write a quick sort algorithm in python."} ] inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt").to(model.device) # tokenizer.eos_token_id is the id of <|EOT|> token outputs = model.generate(inputs, max_new_tokens=512, do_sample=False, top_k=50, top_p=0.95, num_return_sequences=1, eos_token_id=tokenizer.eos_token_id) print(tokenizer.decode(outputs[0][len(inputs[0]):], skip_special_tokens=True)) ``` ### 4. License This code repository is licensed under the MIT License. The use of DeepSeek Coder models is subject to the Model License. DeepSeek Coder supports commercial use. See the [LICENSE-MODEL](https://github.com/deepseek-ai/deepseek-coder/blob/main/LICENSE-MODEL) for more details. ### 5. Contact If you have any questions, please raise an issue or contact us at [agi_code@deepseek.com](mailto:agi_code@deepseek.com).

facebook/musicgen-small

facebook

transformers

text-to-audio

--- inference: true tags: - musicgen license: cc-by-nc-4.0 pipeline_tag: text-to-audio widget: - text: "a funky house with 80s hip hop vibes" example_title: "Prompt 1" - text: "a chill song with influences from lofi, chillstep and downtempo" example_title: "Prompt 2" - text: "a catchy beat for a podcast intro" example_title: "Prompt 3" --- # MusicGen - Small - 300M MusicGen is a text-to-music model capable of genreating high-quality music samples conditioned on text descriptions or audio prompts. It is a single stage auto-regressive Transformer model trained over a 32kHz EnCodec tokenizer with 4 codebooks sampled at 50 Hz. Unlike existing methods, like MusicLM, MusicGen doesn't require a self-supervised semantic representation, and it generates all 4 codebooks in one pass. By introducing a small delay between the codebooks, we show we can predict them in parallel, thus having only 50 auto-regressive steps per second of audio. MusicGen was published in [Simple and Controllable Music Generation](https://arxiv.org/abs/2306.05284) by *Jade Copet, Felix Kreuk, Itai Gat, Tal Remez, David Kant, Gabriel Synnaeve, Yossi Adi, Alexandre Défossez*. Four checkpoints are released: - [**small** (this checkpoint)](https://huggingface.co/facebook/musicgen-small) - [medium](https://huggingface.co/facebook/musicgen-medium) - [large](https://huggingface.co/facebook/musicgen-large) - [melody](https://huggingface.co/facebook/musicgen-melody) ## Example Try out MusicGen yourself! * Audiocraft Colab: <a target="_blank" href="https://colab.research.google.com/drive/1fxGqfg96RBUvGxZ1XXN07s3DthrKUl4-?usp=sharing"> <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/> </a> * Hugging Face Colab: <a target="_blank" href="https://colab.research.google.com/github/sanchit-gandhi/notebooks/blob/main/MusicGen.ipynb"> <img src="https://colab.research.google.com/assets/colab-badge.svg" alt="Open In Colab"/> </a> * Hugging Face Demo: <a target="_blank" href="https://huggingface.co/spaces/facebook/MusicGen"> <img src="https://huggingface.co/datasets/huggingface/badges/raw/main/open-in-hf-spaces-sm.svg" alt="Open in HuggingFace"/> </a> ## 🤗 Transformers Usage You can run MusicGen locally with the 🤗 Transformers library from version 4.31.0 onwards. 1. First install the 🤗 [Transformers library](https://github.com/huggingface/transformers) and scipy: ``` pip install --upgrade pip pip install --upgrade transformers scipy ``` 2. Run inference via the `Text-to-Audio` (TTA) pipeline. You can infer the MusicGen model via the TTA pipeline in just a few lines of code! ```python from transformers import pipeline import scipy synthesiser = pipeline("text-to-audio", "facebook/musicgen-small") music = synthesiser("lo-fi music with a soothing melody", forward_params={"do_sample": True}) scipy.io.wavfile.write("musicgen_out.wav", rate=music["sampling_rate"], data=music["audio"]) ``` 3. Run inference via the Transformers modelling code. You can use the processor + generate code to convert text into a mono 32 kHz audio waveform for more fine-grained control. ```python from transformers import AutoProcessor, MusicgenForConditionalGeneration processor = AutoProcessor.from_pretrained("facebook/musicgen-small") model = MusicgenForConditionalGeneration.from_pretrained("facebook/musicgen-small") inputs = processor( text=["80s pop track with bassy drums and synth", "90s rock song with loud guitars and heavy drums"], padding=True, return_tensors="pt", ) audio_values = model.generate(**inputs, max_new_tokens=256) ``` 3. Listen to the audio samples either in an ipynb notebook: ```python from IPython.display import Audio sampling_rate = model.config.audio_encoder.sampling_rate Audio(audio_values[0].numpy(), rate=sampling_rate) ``` Or save them as a `.wav` file using a third-party library, e.g. `scipy`: ```python import scipy sampling_rate = model.config.audio_encoder.sampling_rate scipy.io.wavfile.write("musicgen_out.wav", rate=sampling_rate, data=audio_values[0, 0].numpy()) ``` For more details on using the MusicGen model for inference using the 🤗 Transformers library, refer to the [MusicGen docs](https://huggingface.co/docs/transformers/model_doc/musicgen). ## Audiocraft Usage You can also run MusicGen locally through the original [Audiocraft library]((https://github.com/facebookresearch/audiocraft): 1. First install the [`audiocraft` library](https://github.com/facebookresearch/audiocraft) ``` pip install git+https://github.com/facebookresearch/audiocraft.git ``` 2. Make sure to have [`ffmpeg`](https://ffmpeg.org/download.html) installed: ``` apt-get install ffmpeg ``` 3. Run the following Python code: ```py from audiocraft.models import MusicGen from audiocraft.data.audio import audio_write model = MusicGen.get_pretrained("small") model.set_generation_params(duration=8) # generate 8 seconds. descriptions = ["happy rock", "energetic EDM"] wav = model.generate(descriptions) # generates 2 samples. for idx, one_wav in enumerate(wav): # Will save under {idx}.wav, with loudness normalization at -14 db LUFS. audio_write(f'{idx}', one_wav.cpu(), model.sample_rate, strategy="loudness") ``` ## Model details **Organization developing the model:** The FAIR team of Meta AI. **Model date:** MusicGen was trained between April 2023 and May 2023. **Model version:** This is the version 1 of the model. **Model type:** MusicGen consists of an EnCodec model for audio tokenization, an auto-regressive language model based on the transformer architecture for music modeling. The model comes in different sizes: 300M, 1.5B and 3.3B parameters ; and two variants: a model trained for text-to-music generation task and a model trained for melody-guided music generation. **Paper or resources for more information:** More information can be found in the paper [Simple and Controllable Music Generation](https://arxiv.org/abs/2306.05284). **Citation details:** ``` @misc{copet2023simple, title={Simple and Controllable Music Generation}, author={Jade Copet and Felix Kreuk and Itai Gat and Tal Remez and David Kant and Gabriel Synnaeve and Yossi Adi and Alexandre Défossez}, year={2023}, eprint={2306.05284}, archivePrefix={arXiv}, primaryClass={cs.SD} } ``` **License:** Code is released under MIT, model weights are released under CC-BY-NC 4.0. **Where to send questions or comments about the model:** Questions and comments about MusicGen can be sent via the [Github repository](https://github.com/facebookresearch/audiocraft) of the project, or by opening an issue. ## Intended use **Primary intended use:** The primary use of MusicGen is research on AI-based music generation, including: - Research efforts, such as probing and better understanding the limitations of generative models to further improve the state of science - Generation of music guided by text or melody to understand current abilities of generative AI models by machine learning amateurs **Primary intended users:** The primary intended users of the model are researchers in audio, machine learning and artificial intelligence, as well as amateur seeking to better understand those models. **Out-of-scope use cases:** The model should not be used on downstream applications without further risk evaluation and mitigation. The model should not be used to intentionally create or disseminate music pieces that create hostile or alienating environments for people. This includes generating music that people would foreseeably find disturbing, distressing, or offensive; or content that propagates historical or current stereotypes. ## Metrics **Models performance measures:** We used the following objective measure to evaluate the model on a standard music benchmark: - Frechet Audio Distance computed on features extracted from a pre-trained audio classifier (VGGish) - Kullback-Leibler Divergence on label distributions extracted from a pre-trained audio classifier (PaSST) - CLAP Score between audio embedding and text embedding extracted from a pre-trained CLAP model Additionally, we run qualitative studies with human participants, evaluating the performance of the model with the following axes: - Overall quality of the music samples; - Text relevance to the provided text input; - Adherence to the melody for melody-guided music generation. More details on performance measures and human studies can be found in the paper. **Decision thresholds:** Not applicable. ## Evaluation datasets The model was evaluated on the [MusicCaps benchmark](https://www.kaggle.com/datasets/googleai/musiccaps) and on an in-domain held-out evaluation set, with no artist overlap with the training set. ## Training datasets The model was trained on licensed data using the following sources: the [Meta Music Initiative Sound Collection](https://www.fb.com/sound), [Shutterstock music collection](https://www.shutterstock.com/music) and the [Pond5 music collection](https://www.pond5.com/). See the paper for more details about the training set and corresponding preprocessing. ## Evaluation results Below are the objective metrics obtained on MusicCaps with the released model. Note that for the publicly released models, we had all the datasets go through a state-of-the-art music source separation method, namely using the open source [Hybrid Transformer for Music Source Separation](https://github.com/facebookresearch/demucs) (HT-Demucs), in order to keep only the instrumental part. This explains the difference in objective metrics with the models used in the paper. | Model | Frechet Audio Distance | KLD | Text Consistency | Chroma Cosine Similarity | |---|---|---|---|---| | **facebook/musicgen-small** | 4.88 | 1.42 | 0.27 | - | | facebook/musicgen-medium | 5.14 | 1.38 | 0.28 | - | | facebook/musicgen-large | 5.48 | 1.37 | 0.28 | - | | facebook/musicgen-melody | 4.93 | 1.41 | 0.27 | 0.44 | More information can be found in the paper [Simple and Controllable Music Generation](https://arxiv.org/abs/2306.05284), in the Results section. ## Limitations and biases **Data:** The data sources used to train the model are created by music professionals and covered by legal agreements with the right holders. The model is trained on 20K hours of data, we believe that scaling the model on larger datasets can further improve the performance of the model. **Mitigations:** Vocals have been removed from the data source using corresponding tags, and then using a state-of-the-art music source separation method, namely using the open source [Hybrid Transformer for Music Source Separation](https://github.com/facebookresearch/demucs) (HT-Demucs). **Limitations:** - The model is not able to generate realistic vocals. - The model has been trained with English descriptions and will not perform as well in other languages. - The model does not perform equally well for all music styles and cultures. - The model sometimes generates end of songs, collapsing to silence. - It is sometimes difficult to assess what types of text descriptions provide the best generations. Prompt engineering may be required to obtain satisfying results. **Biases:** The source of data is potentially lacking diversity and all music cultures are not equally represented in the dataset. The model may not perform equally well on the wide variety of music genres that exists. The generated samples from the model will reflect the biases from the training data. Further work on this model should include methods for balanced and just representations of cultures, for example, by scaling the training data to be both diverse and inclusive. **Risks and harms:** Biases and limitations of the model may lead to generation of samples that may be considered as biased, inappropriate or offensive. We believe that providing the code to reproduce the research and train new models will allow to broaden the application to new and more representative data. **Use cases:** Users must be aware of the biases, limitations and risks of the model. MusicGen is a model developed for artificial intelligence research on controllable music generation. As such, it should not be used for downstream applications without further investigation and mitigation of risks.

timm/pit_b_224.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for pit_b_224.in1k A PiT (Pooling based Vision Transformer) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 73.8 - GMACs: 12.4 - Activations (M): 32.9 - Image size: 224 x 224 - **Papers:** - Rethinking Spatial Dimensions of Vision Transformers: https://arxiv.org/abs/2103.16302 - **Dataset:** ImageNet-1k - **Original:** https://github.com/naver-ai/pit ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('pit_b_224.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'pit_b_224.in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 256, 31, 31]) # torch.Size([1, 512, 16, 16]) # torch.Size([1, 1024, 8, 8]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'pit_b_224.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1, 1024) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @inproceedings{heo2021pit, title={Rethinking Spatial Dimensions of Vision Transformers}, author={Byeongho Heo and Sangdoo Yun and Dongyoon Han and Sanghyuk Chun and Junsuk Choe and Seong Joon Oh}, booktitle = {International Conference on Computer Vision (ICCV)}, year={2021}, } ```

timm/res2next50.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: unknown datasets: - imagenet-1k --- # Model card for res2next50.in1k A Res2Net (Multi-Scale ResNet) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 24.7 - GMACs: 4.2 - Activations (M): 13.7 - Image size: 224 x 224 - **Papers:** - Res2Net: A New Multi-scale Backbone Architecture: https://arxiv.org/abs/1904.01169 - **Dataset:** ImageNet-1k - **Original:** https://github.com/gasvn/Res2Net/ ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('res2next50.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'res2next50.in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 112, 112]) # torch.Size([1, 256, 56, 56]) # torch.Size([1, 512, 28, 28]) # torch.Size([1, 1024, 14, 14]) # torch.Size([1, 2048, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'res2next50.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{gao2019res2net, title={Res2Net: A New Multi-scale Backbone Architecture}, author={Gao, Shang-Hua and Cheng, Ming-Ming and Zhao, Kai and Zhang, Xin-Yu and Yang, Ming-Hsuan and Torr, Philip}, journal={IEEE TPAMI}, doi={10.1109/TPAMI.2019.2938758}, } ```

timm/res2net101_26w_4s.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: unknown datasets: - imagenet-1k --- # Model card for res2net101_26w_4s.in1k A Res2Net (Multi-Scale ResNet) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 45.2 - GMACs: 8.1 - Activations (M): 18.4 - Image size: 224 x 224 - **Papers:** - Res2Net: A New Multi-scale Backbone Architecture: https://arxiv.org/abs/1904.01169 - **Dataset:** ImageNet-1k - **Original:** https://github.com/gasvn/Res2Net/ ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('res2net101_26w_4s.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'res2net101_26w_4s.in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 112, 112]) # torch.Size([1, 256, 56, 56]) # torch.Size([1, 512, 28, 28]) # torch.Size([1, 1024, 14, 14]) # torch.Size([1, 2048, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'res2net101_26w_4s.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{gao2019res2net, title={Res2Net: A New Multi-scale Backbone Architecture}, author={Gao, Shang-Hua and Cheng, Ming-Ming and Zhao, Kai and Zhang, Xin-Yu and Yang, Ming-Hsuan and Torr, Philip}, journal={IEEE TPAMI}, doi={10.1109/TPAMI.2019.2938758}, } ```

nvidia/Llama3-ChatQA-1.5-8B

nvidia

transformers

text-generation

--- license: llama3 language: - en pipeline_tag: text-generation tags: - nvidia - chatqa-1.5 - chatqa - llama-3 - pytorch --- ## Model Details We introduce Llama3-ChatQA-1.5, which excels at conversational question answering (QA) and retrieval-augmented generation (RAG). Llama3-ChatQA-1.5 is developed using an improved training recipe from [ChatQA paper](https://arxiv.org/pdf/2401.10225), and it is built on top of [Llama-3 base model](https://huggingface.co/meta-llama/Meta-Llama-3-8B). Specifically, we incorporate more conversational QA data to enhance its tabular and arithmetic calculation capability. Llama3-ChatQA-1.5 has two variants: Llama3-ChatQA-1.5-8B and Llama3-ChatQA-1.5-70B. Both models were originally trained using [Megatron-LM](https://github.com/NVIDIA/Megatron-LM), we converted the checkpoints to Hugging Face format. **For more information about ChatQA, check the [website](https://chatqa-project.github.io/)!** ## Other Resources [Llama3-ChatQA-1.5-70B](https://huggingface.co/nvidia/Llama3-ChatQA-1.5-70B) &ensp; [Evaluation Data](https://huggingface.co/datasets/nvidia/ChatRAG-Bench) &ensp; [Training Data](https://huggingface.co/datasets/nvidia/ChatQA-Training-Data) &ensp; [Retriever](https://huggingface.co/nvidia/dragon-multiturn-query-encoder) &ensp; [Website](https://chatqa-project.github.io/) &ensp; [Paper](https://arxiv.org/pdf/2401.10225) ## Benchmark Results Results in [ChatRAG Bench](https://huggingface.co/datasets/nvidia/ChatRAG-Bench) are as follows: | | ChatQA-1.0-7B | Command-R-Plus | Llama3-instruct-70b | GPT-4-0613 | GPT-4-Turbo | ChatQA-1.0-70B | ChatQA-1.5-8B | ChatQA-1.5-70B | | -- |:--:|:--:|:--:|:--:|:--:|:--:|:--:|:--:| | Doc2Dial | 37.88 | 33.51 | 37.88 | 34.16 | 35.35 | 38.90 | 39.33 | 41.26 | | QuAC | 29.69 | 34.16 | 36.96 | 40.29 | 40.10 | 41.82 | 39.73 | 38.82 | | QReCC | 46.97 | 49.77 | 51.34 | 52.01 | 51.46 | 48.05 | 49.03 | 51.40 | | CoQA | 76.61 | 69.71 | 76.98 | 77.42 | 77.73 | 78.57 | 76.46 | 78.44 | | DoQA | 41.57 | 40.67 | 41.24 | 43.39 | 41.60 | 51.94 | 49.60 | 50.67 | | ConvFinQA | 51.61 | 71.21 | 76.6 | 81.28 | 84.16 | 73.69 | 78.46 | 81.88 | | SQA | 61.87 | 74.07 | 69.61 | 79.21 | 79.98 | 69.14 | 73.28 | 83.82 | | TopioCQA | 45.45 | 53.77 | 49.72 | 45.09 | 48.32 | 50.98 | 49.96 | 55.63 | | HybriDial* | 54.51 | 46.7 | 48.59 | 49.81 | 47.86 | 56.44 | 65.76 | 68.27 | | INSCIT | 30.96 | 35.76 | 36.23 | 36.34 | 33.75 | 31.90 | 30.10 | 32.31 | | Average (all) | 47.71 | 50.93 | 52.52 | 53.90 | 54.03 | 54.14 | 55.17 | 58.25 | | Average (exclude HybriDial) | 46.96 | 51.40 | 52.95 | 54.35 | 54.72 | 53.89 | 53.99 | 57.14 | Note that ChatQA-1.5 is built based on Llama-3 base model, and ChatQA-1.0 is built based on Llama-2 base model. ChatQA-1.5 models use HybriDial training dataset. To ensure fair comparison, we also compare average scores excluding HybriDial. The data and evaluation scripts for ChatRAG Bench can be found [here](https://huggingface.co/datasets/nvidia/ChatRAG-Bench). ## Prompt Format **We highly recommend that you use the prompt format we provide, as follows:** ### when context is available <pre> System: {System} {Context} User: {Question} Assistant: {Response} User: {Question} Assistant: </pre> ### when context is not available <pre> System: {System} User: {Question} Assistant: {Response} User: {Question} Assistant: </pre> **The content of the system's turn (i.e., {System}) for both scenarios is as follows:** <pre> This is a chat between a user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions based on the context. The assistant should also indicate when the answer cannot be found in the context. </pre> **Note that our ChatQA-1.5 models are optimized for the capability with context, e.g., over documents or retrieved context.** ## How to use ### take the whole document as context This can be applied to the scenario where the whole document can be fitted into the model, so that there is no need to run retrieval over the document. ```python from transformers import AutoTokenizer, AutoModelForCausalLM import torch model_id = "nvidia/Llama3-ChatQA-1.5-8B" tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, device_map="auto") messages = [ {"role": "user", "content": "what is the percentage change of the net income from Q4 FY23 to Q4 FY24?"} ] document = """NVIDIA (NASDAQ: NVDA) today reported revenue for the fourth quarter ended January 28, 2024, of $22.1 billion, up 22% from the previous quarter and up 265% from a year ago.\nFor the quarter, GAAP earnings per diluted share was $4.93, up 33% from the previous quarter and up 765% from a year ago. Non-GAAP earnings per diluted share was $5.16, up 28% from the previous quarter and up 486% from a year ago.\nQ4 Fiscal 2024 Summary\nGAAP\n| $ in millions, except earnings per share | Q4 FY24 | Q3 FY24 | Q4 FY23 | Q/Q | Y/Y |\n| Revenue | $22,103 | $18,120 | $6,051 | Up 22% | Up 265% |\n| Gross margin | 76.0% | 74.0% | 63.3% | Up 2.0 pts | Up 12.7 pts |\n| Operating expenses | $3,176 | $2,983 | $2,576 | Up 6% | Up 23% |\n| Operating income | $13,615 | $10,417 | $1,257 | Up 31% | Up 983% |\n| Net income | $12,285 | $9,243 | $1,414 | Up 33% | Up 769% |\n| Diluted earnings per share | $4.93 | $3.71 | $0.57 | Up 33% | Up 765% |""" def get_formatted_input(messages, context): system = "System: This is a chat between a user and an artificial intelligence assistant. The assistant gives helpful, detailed, and polite answers to the user's questions based on the context. The assistant should also indicate when the answer cannot be found in the context." instruction = "Please give a full and complete answer for the question." for item in messages: if item['role'] == "user": ## only apply this instruction for the first user turn item['content'] = instruction + " " + item['content'] break conversation = '\n\n'.join(["User: " + item["content"] if item["role"] == "user" else "Assistant: " + item["content"] for item in messages]) + "\n\nAssistant:" formatted_input = system + "\n\n" + context + "\n\n" + conversation return formatted_input formatted_input = get_formatted_input(messages, document) tokenized_prompt = tokenizer(tokenizer.bos_token + formatted_input, return_tensors="pt").to(model.device) terminators = [ tokenizer.eos_token_id, tokenizer.convert_tokens_to_ids("<|eot_id|>") ] outputs = model.generate(input_ids=tokenized_prompt.input_ids, attention_mask=tokenized_prompt.attention_mask, max_new_tokens=128, eos_token_id=terminators) response = outputs[0][tokenized_prompt.input_ids.shape[-1]:] print(tokenizer.decode(response, skip_special_tokens=True)) ``` ### run retrieval to get top-n chunks as context This can be applied to the scenario when the document is very long, so that it is necessary to run retrieval. Here, we use our [Dragon-multiturn](https://huggingface.co/nvidia/dragon-multiturn-query-encoder) retriever which can handle conversatinoal query. In addition, we provide a few [documents](https://huggingface.co/nvidia/Llama3-ChatQA-1.5-8B/tree/main/docs) for users to play with. ```python from transformers import AutoTokenizer, AutoModelForCausalLM, AutoModel import torch import json ## load ChatQA-1.5 tokenizer and model model_id = "nvidia/Llama3-ChatQA-1.5-8B" tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, device_map="auto") ## load retriever tokenizer and model retriever_tokenizer = AutoTokenizer.from_pretrained('nvidia/dragon-multiturn-query-encoder') query_encoder = AutoModel.from_pretrained('nvidia/dragon-multiturn-query-encoder') context_encoder = AutoModel.from_pretrained('nvidia/dragon-multiturn-context-encoder') ## prepare documents, we take landrover car manual document that we provide as an example chunk_list = json.load(open("docs.json"))['landrover'] messages = [ {"role": "user", "content": "how to connect the bluetooth in the car?"} ] ### running retrieval ## convert query into a format as follows: ## user: {user}\nagent: {agent}\nuser: {user} formatted_query_for_retriever = '\n'.join([turn['role'] + ": " + turn['content'] for turn in messages]).strip() query_input = retriever_tokenizer(formatted_query_for_retriever, return_tensors='pt') ctx_input = retriever_tokenizer(chunk_list, padding=True, truncation=True, max_length=512, return_tensors='pt') query_emb = query_encoder(**query_input).last_hidden_state[:, 0, :] ctx_emb = context_encoder(**ctx_input).last_hidden_state[:, 0, :] ## Compute similarity scores using dot product and rank the similarity similarities = query_emb.matmul(ctx_emb.transpose(0, 1)) # (1, num_ctx) ranked_results = torch.argsort(similarities, dim=-1, descending=True) # (1, num_ctx) ## get top-n chunks (n=5) retrieved_chunks = [chunk_list[idx] for idx in ranked_results.tolist()[0][:5]] context = "\n\n".join(retrieved_chunks) ### running text generation formatted_input = get_formatted_input(messages, context) tokenized_prompt = tokenizer(tokenizer.bos_token + formatted_input, return_tensors="pt").to(model.device) terminators = [ tokenizer.eos_token_id, tokenizer.convert_tokens_to_ids("<|eot_id|>") ] outputs = model.generate(input_ids=tokenized_prompt.input_ids, attention_mask=tokenized_prompt.attention_mask, max_new_tokens=128, eos_token_id=terminators) response = outputs[0][tokenized_prompt.input_ids.shape[-1]:] print(tokenizer.decode(response, skip_special_tokens=True)) ``` ## Correspondence to Zihan Liu (zihanl@nvidia.com), Wei Ping (wping@nvidia.com) ## Citation <pre> @article{liu2024chatqa, title={ChatQA: Surpassing GPT-4 on Conversational QA and RAG}, author={Liu, Zihan and Ping, Wei and Roy, Rajarshi and Xu, Peng and Lee, Chankyu and Shoeybi, Mohammad and Catanzaro, Bryan}, journal={arXiv preprint arXiv:2401.10225}, year={2024}} </pre> ## License The use of this model is governed by the [META LLAMA 3 COMMUNITY LICENSE AGREEMENT](https://llama.meta.com/llama3/license/)

timm/mobilenetv3_large_100.miil_in21k_ft_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k - imagenet-21k-p --- # Model card for mobilenetv3_large_100.miil_in21k_ft_in1k A MobileNet-v3 image classification model. Petrained on ImageNet-21k-P and fine-tuned on ImageNet-1k by Alibaba MIIL. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 5.5 - GMACs: 0.2 - Activations (M): 4.4 - Image size: 224 x 224 - **Papers:** - Searching for MobileNetV3: https://arxiv.org/abs/1905.02244 - **Dataset:** ImageNet-1k - **Pretrain Dataset:** ImageNet-21k-P ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('mobilenetv3_large_100.miil_in21k_ft_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mobilenetv3_large_100.miil_in21k_ft_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 16, 112, 112]) # torch.Size([1, 24, 56, 56]) # torch.Size([1, 40, 28, 28]) # torch.Size([1, 112, 14, 14]) # torch.Size([1, 960, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mobilenetv3_large_100.miil_in21k_ft_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 960, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @inproceedings{howard2019searching, title={Searching for mobilenetv3}, author={Howard, Andrew and Sandler, Mark and Chu, Grace and Chen, Liang-Chieh and Chen, Bo and Tan, Mingxing and Wang, Weijun and Zhu, Yukun and Pang, Ruoming and Vasudevan, Vijay and others}, booktitle={Proceedings of the IEEE/CVF international conference on computer vision}, pages={1314--1324}, year={2019} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

katuni4ka/tiny-random-minicpm

katuni4ka

transformers

text-generation

Entry not found

google/gemma-1.1-2b-it

google

transformers

text-generation

--- library_name: transformers license: gemma widget: - messages: - role: user content: How does the brain work? inference: parameters: max_new_tokens: 200 extra_gated_heading: Access Gemma on Hugging Face extra_gated_prompt: To access Gemma on Hugging Face, you’re required to review and agree to Google’s usage license. To do this, please ensure you’re logged-in to Hugging Face and click below. Requests are processed immediately. extra_gated_button_content: Acknowledge license --- # Gemma Model Card **Model Page**: [Gemma](https://ai.google.dev/gemma/docs) This model card corresponds to the latest 2B instruct version of the Gemma model. Here you can find other models in the Gemma family: | | Base | Instruct | |----|----------------------------------------------------|----------------------------------------------------------------------| | 2B | [gemma-2b](https://huggingface.co/google/gemma-2b) | [**gemma-1.1-2b-it**](https://huggingface.co/google/gemma-1.1-2b-it) | | 7B | [gemma-7b](https://huggingface.co/google/gemma-7b) | [gemma-1.1-7b-it](https://huggingface.co/google/gemma-1.1-7b-it) | **Release Notes** This is Gemma 1.1 2B (IT), an update over the original instruction-tuned Gemma release. Gemma 1.1 was trained using a novel RLHF method, leading to substantial gains on quality, coding capabilities, factuality, instruction following and multi-turn conversation quality. We also fixed a bug in multi-turn conversations, and made sure that model responses don't always start with `"Sure,"`. We believe this release represents an improvement for most use cases, but we encourage users to test in their particular applications. The previous model [will continue to be available in the same repo](https://huggingface.co/google/gemma-2b-it). We appreciate the enthusiastic adoption of Gemma, and we continue to welcome all feedback from the community. **Resources and Technical Documentation**: * [Responsible Generative AI Toolkit](https://ai.google.dev/responsible) * [Gemma on Kaggle](https://www.kaggle.com/models/google/gemma) * [Gemma on Vertex Model Garden](https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/335) **Terms of Use**: [Terms](https://www.kaggle.com/models/google/gemma/license/consent/verify/huggingface?returnModelRepoId=google/gemma-1.1-2b-it) **Authors**: Google ## Model Information Summary description and brief definition of inputs and outputs. ### Description Gemma is a family of lightweight, state-of-the-art open models from Google, built from the same research and technology used to create the Gemini models. They are text-to-text, decoder-only large language models, available in English, with open weights, pre-trained variants, and instruction-tuned variants. Gemma models are well-suited for a variety of text generation tasks, including question answering, summarization, and reasoning. Their relatively small size makes it possible to deploy them in environments with limited resources such as a laptop, desktop or your own cloud infrastructure, democratizing access to state of the art AI models and helping foster innovation for everyone. ### Usage Below we share some code snippets on how to get quickly started with running the model. First make sure to `pip install -U transformers`, then copy the snippet from the section that is relevant for your usecase. #### Running the model on a CPU As explained below, we recommend `torch.bfloat16` as the default dtype. You can use [a different precision](#precisions) if necessary. ```python from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("google/gemma-1.1-2b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-1.1-2b-it", torch_dtype=torch.bfloat16 ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt") outputs = model.generate(**input_ids, max_new_tokens=50) print(tokenizer.decode(outputs[0])) ``` #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("google/gemma-1.1-2b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-1.1-2b-it", device_map="auto", torch_dtype=torch.bfloat16 ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` <a name="precisions"></a> #### Running the model on a GPU using different precisions The native weights of this model were exported in `bfloat16` precision. You can use `float16`, which may be faster on certain hardware, indicating the `torch_dtype` when loading the model. For convenience, the `float16` revision of the repo contains a copy of the weights already converted to that precision. You can also use `float32` if you skip the dtype, but no precision increase will occur (model weights will just be upcasted to `float32`). See examples below. * _Using `torch.float16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("google/gemma-1.1-2b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-1.1-2b-it", device_map="auto", torch_dtype=torch.float16, revision="float16", ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using `torch.bfloat16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-2b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-1.1-2b-it", device_map="auto", torch_dtype=torch.bfloat16 ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Upcasting to `torch.float32`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-1.1-2b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-1.1-2b-it", device_map="auto" ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` * _Using 8-bit precision (int8)_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-1.1-2b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-1.1-2b-it", quantization_config=quantization_config ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using 4-bit precision_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-1.1-2b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-1.1-2b-it", quantization_config=quantization_config ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Other optimizations * _Flash Attention 2_ First make sure to install `flash-attn` in your environment `pip install flash-attn` ```diff model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, + attn_implementation="flash_attention_2" ).to(0) ``` #### Running the model in JAX / Flax Use the `flax` branch of the repository: ```python import jax.numpy as jnp from transformers import AutoTokenizer, FlaxGemmaForCausalLM model_id = "google/gemma-1.1-2b-it" tokenizer = AutoTokenizer.from_pretrained(model_id) tokenizer.padding_side = "left" model, params = FlaxGemmaForCausalLM.from_pretrained( model_id, dtype=jnp.bfloat16, revision="flax", _do_init=False, ) inputs = tokenizer("Valencia and Málaga are", return_tensors="np", padding=True) output = model.generate(**inputs, params=params, max_new_tokens=20, do_sample=False) output_text = tokenizer.batch_decode(output.sequences, skip_special_tokens=True) ``` [Check this notebook](https://colab.research.google.com/github/sanchit-gandhi/notebooks/blob/main/jax_gemma.ipynb) for a comprehensive walkthrough on how to parallelize JAX inference. ### Chat Template The instruction-tuned models use a chat template that must be adhered to for conversational use. The easiest way to apply it is using the tokenizer's built-in chat template, as shown in the following snippet. Let's load the model and apply the chat template to a conversation. In this example, we'll start with a single user interaction: ```py from transformers import AutoTokenizer, AutoModelForCausalLM import transformers import torch model_id = "google/gemma-1.1-2b-it" dtype = torch.bfloat16 tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained( model_id, device_map="cuda", torch_dtype=dtype, ) chat = [ { "role": "user", "content": "Write a hello world program" }, ] prompt = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True) ``` At this point, the prompt contains the following text: ``` <bos><start_of_turn>user Write a hello world program<end_of_turn> <start_of_turn>model ``` As you can see, each turn is preceded by a `<start_of_turn>` delimiter and then the role of the entity (either `user`, for content supplied by the user, or `model` for LLM responses). Turns finish with the `<end_of_turn>` token. You can follow this format to build the prompt manually, if you need to do it without the tokenizer's chat template. After the prompt is ready, generation can be performed like this: ```py inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt") outputs = model.generate(input_ids=inputs.to(model.device), max_new_tokens=150) ``` ### Fine-tuning You can find some fine-tuning scripts under the [`examples/` directory](https://huggingface.co/google/gemma-7b/tree/main/examples) of [`google/gemma-7b`](https://huggingface.co/google/gemma-7b) repository. To adapt them to this model, simply change the model-id to `google/gemma-1.1-2b-it`. We provide: * A script to perform Supervised Fine-Tuning (SFT) on UltraChat dataset using QLoRA * A script to perform SFT using FSDP on TPU devices * A notebook that you can run on a free-tier Google Colab instance to perform SFT on the English quotes dataset ### Inputs and outputs * **Input:** Text string, such as a question, a prompt, or a document to be summarized. * **Output:** Generated English-language text in response to the input, such as an answer to a question, or a summary of a document. ## Model Data Data used for model training and how the data was processed. ### Training Dataset These models were trained on a dataset of text data that includes a wide variety of sources, totaling 6 trillion tokens. Here are the key components: * Web Documents: A diverse collection of web text ensures the model is exposed to a broad range of linguistic styles, topics, and vocabulary. Primarily English-language content. * Code: Exposing the model to code helps it to learn the syntax and patterns of programming languages, which improves its ability to generate code or understand code-related questions. * Mathematics: Training on mathematical text helps the model learn logical reasoning, symbolic representation, and to address mathematical queries. The combination of these diverse data sources is crucial for training a powerful language model that can handle a wide variety of different tasks and text formats. ### Data Preprocessing Here are the key data cleaning and filtering methods applied to the training data: * CSAM Filtering: Rigorous CSAM (Child Sexual Abuse Material) filtering was applied at multiple stages in the data preparation process to ensure the exclusion of harmful and illegal content * Sensitive Data Filtering: As part of making Gemma pre-trained models safe and reliable, automated techniques were used to filter out certain personal information and other sensitive data from training sets. * Additional methods: Filtering based on content quality and safely in line with [our policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11). ## Implementation Information Details about the model internals. ### Hardware Gemma was trained using the latest generation of [Tensor Processing Unit (TPU)](https://cloud.google.com/tpu/docs/intro-to-tpu) hardware (TPUv5e). Training large language models requires significant computational power. TPUs, designed specifically for matrix operations common in machine learning, offer several advantages in this domain: * Performance: TPUs are specifically designed to handle the massive computations involved in training LLMs. They can speed up training considerably compared to CPUs. * Memory: TPUs often come with large amounts of high-bandwidth memory, allowing for the handling of large models and batch sizes during training. This can lead to better model quality. * Scalability: TPU Pods (large clusters of TPUs) provide a scalable solution for handling the growing complexity of large foundation models. You can distribute training across multiple TPU devices for faster and more efficient processing. * Cost-effectiveness: In many scenarios, TPUs can provide a more cost-effective solution for training large models compared to CPU-based infrastructure, especially when considering the time and resources saved due to faster training. * These advantages are aligned with [Google's commitments to operate sustainably](https://sustainability.google/operating-sustainably/). ### Software Training was done using [JAX](https://github.com/google/jax) and [ML Pathways](https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture/ml-pathways). JAX allows researchers to take advantage of the latest generation of hardware, including TPUs, for faster and more efficient training of large models. ML Pathways is Google's latest effort to build artificially intelligent systems capable of generalizing across multiple tasks. This is specially suitable for [foundation models](https://ai.google/discover/foundation-models/), including large language models like these ones. Together, JAX and ML Pathways are used as described in the [paper about the Gemini family of models](https://arxiv.org/abs/2312.11805); "the 'single controller' programming model of Jax and Pathways allows a single Python process to orchestrate the entire training run, dramatically simplifying the development workflow." ## Evaluation Model evaluation metrics and results. ### Benchmark Results The pre-trained base models were evaluated against a large collection of different datasets and metrics to cover different aspects of text generation: | Benchmark | Metric | 2B Params | 7B Params | | ------------------------------ | ------------- | ----------- | --------- | | [MMLU](https://arxiv.org/abs/2009.03300) | 5-shot, top-1 | 42.3 | 64.3 | | [HellaSwag](https://arxiv.org/abs/1905.07830) | 0-shot |71.4 | 81.2 | | [PIQA](https://arxiv.org/abs/1911.11641) | 0-shot | 77.3 | 81.2 | | [SocialIQA](https://arxiv.org/abs/1904.09728) | 0-shot | 49.7 | 51.8 | | [BooIQ](https://arxiv.org/abs/1905.10044) | 0-shot | 69.4 | 83.2 | | [WinoGrande](https://arxiv.org/abs/1907.10641) | partial score | 65.4 | 72.3 | | [CommonsenseQA](https://arxiv.org/abs/1811.00937) | 7-shot | 65.3 | 71.3 | | [OpenBookQA](https://arxiv.org/abs/1809.02789) | | 47.8 | 52.8 | | [ARC-e](https://arxiv.org/abs/1911.01547) | | 73.2 | 81.5 | | [ARC-c](https://arxiv.org/abs/1911.01547) | | 42.1 | 53.2 | | [TriviaQA](https://arxiv.org/abs/1705.03551) | 5-shot | 53.2 | 63.4 | | [Natural Questions](https://github.com/google-research-datasets/natural-questions) | 5-shot | 12.5 | 23 | | [HumanEval](https://arxiv.org/abs/2107.03374) | pass@1 | 22.0 | 32.3 | | [MBPP](https://arxiv.org/abs/2108.07732) | 3-shot | 29.2 | 44.4 | | [GSM8K](https://arxiv.org/abs/2110.14168) | maj@1 | 17.7 | 46.4 | | [MATH](https://arxiv.org/abs/2108.07732) | 4-shot | 11.8 | 24.3 | | [AGIEval](https://arxiv.org/abs/2304.06364) | | 24.2 | 41.7 | | [BIG-Bench](https://arxiv.org/abs/2206.04615) | | 35.2 | 55.1 | | ------------------------------ | ------------- | ----------- | --------- | | **Average** | | **45.0** | **56.9** | ## Ethics and Safety Ethics and safety evaluation approach and results. ### Evaluation Approach Our evaluation methods include structured evaluations and internal red-teaming testing of relevant content policies. Red-teaming was conducted by a number of different teams, each with different goals and human evaluation metrics. These models were evaluated against a number of different categories relevant to ethics and safety, including: * Text-to-Text Content Safety: Human evaluation on prompts covering safety policies including child sexual abuse and exploitation, harassment, violence and gore, and hate speech. * Text-to-Text Representational Harms: Benchmark against relevant academic datasets such as [WinoBias](https://arxiv.org/abs/1804.06876) and [BBQ Dataset](https://arxiv.org/abs/2110.08193v2). * Memorization: Automated evaluation of memorization of training data, including the risk of personally identifiable information exposure. * Large-scale harm: Tests for "dangerous capabilities," such as chemical, biological, radiological, and nuclear (CBRN) risks. ### Evaluation Results The results of ethics and safety evaluations are within acceptable thresholds for meeting [internal policies](https://storage.googleapis.com/gweb-uniblog-publish-prod/documents/2023_Google_AI_Principles_Progress_Update.pdf#page=11) for categories such as child safety, content safety, representational harms, memorization, large-scale harms. On top of robust internal evaluations, the results of well known safety benchmarks like BBQ, BOLD, Winogender, Winobias, RealToxicity, and TruthfulQA are shown here. #### Gemma 1.0 | Benchmark | Metric | Gemma 1.0 IT 2B | Gemma 1.0 IT 7B | | ------------------------ | ------------- | --------------- | --------------- | | [RealToxicity][realtox] | average | 6.86 | 7.90 | | [BOLD][bold] | | 45.57 | 49.08 | | [CrowS-Pairs][crows] | top-1 | 45.82 | 51.33 | | [BBQ Ambig][bbq] | 1-shot, top-1 | 62.58 | 92.54 | | [BBQ Disambig][bbq] | top-1 | 54.62 | 71.99 | | [Winogender][winogender] | top-1 | 51.25 | 54.17 | | [TruthfulQA][truthfulqa] | | 44.84 | 31.81 | | [Winobias 1_2][winobias] | | 56.12 | 59.09 | | [Winobias 2_2][winobias] | | 91.10 | 92.23 | | [Toxigen][toxigen] | | 29.77 | 39.59 | | ------------------------ | ------------- | --------------- | --------------- | #### Gemma 1.1 | Benchmark | Metric | Gemma 1.1 IT 2B | Gemma 1.1 IT 7B | | ------------------------ | ------------- | --------------- | --------------- | | [RealToxicity][realtox] | average | 7.03 | 8.04 | | [BOLD][bold] | | 47.76 | | | [CrowS-Pairs][crows] | top-1 | 45.89 | 49.67 | | [BBQ Ambig][bbq] | 1-shot, top-1 | 58.97 | 86.06 | | [BBQ Disambig][bbq] | top-1 | 53.90 | 85.08 | | [Winogender][winogender] | top-1 | 50.14 | 57.64 | | [TruthfulQA][truthfulqa] | | 44.24 | 45.34 | | [Winobias 1_2][winobias] | | 55.93 | 59.22 | | [Winobias 2_2][winobias] | | 89.46 | 89.2 | | [Toxigen][toxigen] | | 29.64 | 38.75 | | ------------------------ | ------------- | --------------- | --------------- | ## Usage and Limitations These models have certain limitations that users should be aware of. ### Intended Usage Open Large Language Models (LLMs) have a wide range of applications across various industries and domains. The following list of potential uses is not comprehensive. The purpose of this list is to provide contextual information about the possible use-cases that the model creators considered as part of model training and development. * Content Creation and Communication * Text Generation: These models can be used to generate creative text formats such as poems, scripts, code, marketing copy, and email drafts. * Chatbots and Conversational AI: Power conversational interfaces for customer service, virtual assistants, or interactive applications. * Text Summarization: Generate concise summaries of a text corpus, research papers, or reports. * Research and Education * Natural Language Processing (NLP) Research: These models can serve as a foundation for researchers to experiment with NLP techniques, develop algorithms, and contribute to the advancement of the field. * Language Learning Tools: Support interactive language learning experiences, aiding in grammar correction or providing writing practice. * Knowledge Exploration: Assist researchers in exploring large bodies of text by generating summaries or answering questions about specific topics. ### Limitations * Training Data * The quality and diversity of the training data significantly influence the model's capabilities. Biases or gaps in the training data can lead to limitations in the model's responses. * The scope of the training dataset determines the subject areas the model can handle effectively. * Context and Task Complexity * LLMs are better at tasks that can be framed with clear prompts and instructions. Open-ended or highly complex tasks might be challenging. * A model's performance can be influenced by the amount of context provided (longer context generally leads to better outputs, up to a certain point). * Language Ambiguity and Nuance * Natural language is inherently complex. LLMs might struggle to grasp subtle nuances, sarcasm, or figurative language. * Factual Accuracy * LLMs generate responses based on information they learned from their training datasets, but they are not knowledge bases. They may generate incorrect or outdated factual statements. * Common Sense * LLMs rely on statistical patterns in language. They might lack the ability to apply common sense reasoning in certain situations. ### Ethical Considerations and Risks The development of large language models (LLMs) raises several ethical concerns. In creating an open model, we have carefully considered the following: * Bias and Fairness * LLMs trained on large-scale, real-world text data can reflect socio-cultural biases embedded in the training material. These models underwent careful scrutiny, input data pre-processing described and posterior evaluations reported in this card. * Misinformation and Misuse * LLMs can be misused to generate text that is false, misleading, or harmful. * Guidelines are provided for responsible use with the model, see the [Responsible Generative AI Toolkit](http://ai.google.dev/gemma/responsible). * Transparency and Accountability: * This model card summarizes details on the models' architecture, capabilities, limitations, and evaluation processes. * A responsibly developed open model offers the opportunity to share innovation by making LLM technology accessible to developers and researchers across the AI ecosystem. Risks identified and mitigations: * Perpetuation of biases: It's encouraged to perform continuous monitoring (using evaluation metrics, human review) and the exploration of de-biasing techniques during model training, fine-tuning, and other use cases. * Generation of harmful content: Mechanisms and guidelines for content safety are essential. Developers are encouraged to exercise caution and implement appropriate content safety safeguards based on their specific product policies and application use cases. * Misuse for malicious purposes: Technical limitations and developer and end-user education can help mitigate against malicious applications of LLMs. Educational resources and reporting mechanisms for users to flag misuse are provided. Prohibited uses of Gemma models are outlined in the [Gemma Prohibited Use Policy](https://ai.google.dev/gemma/prohibited_use_policy). * Privacy violations: Models were trained on data filtered for removal of PII (Personally Identifiable Information). Developers are encouraged to adhere to privacy regulations with privacy-preserving techniques. ### Benefits At the time of release, this family of models provides high-performance open large language model implementations designed from the ground up for Responsible AI development compared to similarly sized models. Using the benchmark evaluation metrics described in this document, these models have shown to provide superior performance to other, comparably-sized open model alternatives.

timm/mixnet_l.ft_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for mixnet_l.ft_in1k A MixNet image classification model. Fine-tuned on ImageNet-1k from original Tensorflow "SAME" padding weights for use in PyTorch. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 7.3 - GMACs: 0.6 - Activations (M): 10.8 - Image size: 224 x 224 - **Papers:** - MixConv: Mixed Depthwise Convolutional Kernels: https://arxiv.org/abs/1907.09595 - **Dataset:** ImageNet-1k - **Original:** https://github.com/huggingface/pytorch-image-models ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('mixnet_l.ft_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mixnet_l.ft_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 32, 112, 112]) # torch.Size([1, 40, 56, 56]) # torch.Size([1, 56, 28, 28]) # torch.Size([1, 160, 14, 14]) # torch.Size([1, 264, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'mixnet_l.ft_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1536, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ``` ```bibtex @misc{tan2019mixconv, title={MixConv: Mixed Depthwise Convolutional Kernels}, author={Mingxing Tan and Quoc V. Le}, year={2019}, eprint={1907.09595}, archivePrefix={arXiv}, primaryClass={cs.CV} } ```

timm/dla102.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: bsd-3-clause datasets: - imagenet-1k --- # Model card for dla102.in1k A DLA (Deep Layer Aggregation) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 33.3 - GMACs: 7.2 - Activations (M): 14.2 - Image size: 224 x 224 - **Papers:** - Deep Layer Aggregation: https://arxiv.org/abs/1707.06484 - **Original:** https://github.com/ucbdrive/dla - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('dla102.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'dla102.in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 32, 112, 112]) # torch.Size([1, 128, 56, 56]) # torch.Size([1, 256, 28, 28]) # torch.Size([1, 512, 14, 14]) # torch.Size([1, 1024, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'dla102.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1024, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @inproceedings{yu2018deep, title={Deep layer aggregation}, author={Yu, Fisher and Wang, Dequan and Shelhamer, Evan and Darrell, Trevor}, booktitle={Proceedings of the IEEE conference on computer vision and pattern recognition}, year={2018} } ```

katuni4ka/tiny-random-internlm2

katuni4ka

transformers

feature-extraction

Entry not found

duyntnet/gemma-2-27b-it-imatrix-GGUF

duyntnet

transformers

text-generation

--- license: other language: - en pipeline_tag: text-generation inference: false tags: - transformers - gguf - imatrix - gemma-2-27b-it --- Quantizations of https://huggingface.co/google/gemma-2-27b-it **Note**: All quants are created using latest [llama.cpp release](https://github.com/ggerganov/llama.cpp/releases) (b3266). This version (hopefully) fixes all Gemma 2 27B problems. You will need the latest version of llama.cpp to use these quants. # From original readme ### Usage Below we share some code snippets on how to get quickly started with running the model. First make sure to `pip install -U transformers`, then copy the snippet from the section that is relevant for your usecase. #### Running the model on a single / multi GPU ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("google/gemma-2-27b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-2-27b-it", device_map="auto", torch_dtype=torch.bfloat16 ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` <a name="precisions"></a> #### Running the model on a GPU using different precisions The native weights of this model were exported in `bfloat16` precision. You can use `float16`, which may be faster on certain hardware, indicating the `torch_dtype` when loading the model. For convenience, the `float16` revision of the repo contains a copy of the weights already converted to that precision. You can also use `float32` if you skip the dtype, but no precision increase will occur (model weights will just be upcasted to `float32`). See examples below. * _Using `torch.float16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM import torch tokenizer = AutoTokenizer.from_pretrained("google/gemma-2-27b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-2-27b-it", device_map="auto", torch_dtype=torch.float16, revision="float16", ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using `torch.bfloat16`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-2-27b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-2-27b-it", device_map="auto", torch_dtype=torch.bfloat16) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Upcasting to `torch.float32`_ ```python # pip install accelerate from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("google/gemma-2-27b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-2-27b-it", device_map="auto" ) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Quantized Versions through `bitsandbytes` * _Using 8-bit precision (int8)_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_8bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-2-27b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-2-27b-it", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` * _Using 4-bit precision_ ```python # pip install bitsandbytes accelerate from transformers import AutoTokenizer, AutoModelForCausalLM, BitsAndBytesConfig quantization_config = BitsAndBytesConfig(load_in_4bit=True) tokenizer = AutoTokenizer.from_pretrained("google/gemma-2-27b-it") model = AutoModelForCausalLM.from_pretrained( "google/gemma-2-27b-it", quantization_config=quantization_config) input_text = "Write me a poem about Machine Learning." input_ids = tokenizer(input_text, return_tensors="pt").to("cuda") outputs = model.generate(**input_ids) print(tokenizer.decode(outputs[0])) ``` #### Other optimizations * _Flash Attention 2_ First make sure to install `flash-attn` in your environment `pip install flash-attn` ```diff model = AutoModelForCausalLM.from_pretrained( model_id, torch_dtype=torch.float16, + attn_implementation="flash_attention_2" ).to(0) ``` ### Chat Template The instruction-tuned models use a chat template that must be adhered to for conversational use. The easiest way to apply it is using the tokenizer's built-in chat template, as shown in the following snippet. Let's load the model and apply the chat template to a conversation. In this example, we'll start with a single user interaction: ```py from transformers import AutoTokenizer, AutoModelForCausalLM import transformers import torch model_id = "google/gemma-2-27b-it" dtype = torch.bfloat16 tokenizer = AutoTokenizer.from_pretrained(model_id) model = AutoModelForCausalLM.from_pretrained( model_id, device_map="cuda", torch_dtype=dtype, ) chat = [ { "role": "user", "content": "Write a hello world program" }, ] prompt = tokenizer.apply_chat_template(chat, tokenize=False, add_generation_prompt=True) ``` At this point, the prompt contains the following text: ``` <bos><start_of_turn>user Write a hello world program<end_of_turn> <start_of_turn>model ``` As you can see, each turn is preceded by a `<start_of_turn>` delimiter and then the role of the entity (either `user`, for content supplied by the user, or `model` for LLM responses). Turns finish with the `<end_of_turn>` token. You can follow this format to build the prompt manually, if you need to do it without the tokenizer's chat template. After the prompt is ready, generation can be performed like this: ```py inputs = tokenizer.encode(prompt, add_special_tokens=False, return_tensors="pt") outputs = model.generate(input_ids=inputs.to(model.device), max_new_tokens=150) print(tokenizer.decode(outputs[0])) ```

lvwerra/distilbert-imdb

lvwerra

transformers

text-classification

--- license: apache-2.0 tags: - generated_from_trainer datasets: - imdb metrics: - accuracy model-index: - name: distilbert-imdb results: - task: name: Text Classification type: text-classification dataset: name: imdb type: imdb args: plain_text metrics: - name: Accuracy type: accuracy value: 0.928 --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # distilbert-imdb This model is a fine-tuned version of [distilbert-base-uncased](https://huggingface.co/distilbert-base-uncased) on the imdb dataset (training notebook is [here](https://huggingface.co/lvwerra/distilbert-imdb/blob/main/distilbert-imdb-training.ipynb)). It achieves the following results on the evaluation set: - Loss: 0.1903 - Accuracy: 0.928 ## Model description More information needed ## Intended uses & limitations More information needed ## Training and evaluation data More information needed ## Training procedure ### Training hyperparameters The following hyperparameters were used during training: - learning_rate: 5e-05 - train_batch_size: 16 - eval_batch_size: 16 - seed: 42 - optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08 - lr_scheduler_type: linear - num_epochs: 1 ### Training results | Training Loss | Epoch | Step | Validation Loss | Accuracy | |:-------------:|:-----:|:----:|:---------------:|:--------:| | 0.2195 | 1.0 | 1563 | 0.1903 | 0.928 | ### Framework versions - Transformers 4.15.0 - Pytorch 1.10.0+cu111 - Datasets 1.17.0 - Tokenizers 0.10.3

katuni4ka/tiny-random-orion

katuni4ka

transformers

text-generation

Entry not found

timm/poolformer_m36.sail_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for poolformer_m36.sail_in1k A PoolFormer (a MetaFormer) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 56.2 - GMACs: 8.8 - Activations (M): 22.0 - Image size: 224 x 224 - **Papers:** - MetaFormer Is Actually What You Need for Vision: https://arxiv.org/abs/2210.13452 - **Original:** https://github.com/sail-sg/poolformer - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('poolformer_m36.sail_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'poolformer_m36.sail_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 96, 56, 56]) # torch.Size([1, 192, 28, 28]) # torch.Size([1, 384, 14, 14]) # torch.Size([1, 768, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'poolformer_m36.sail_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 768, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @inproceedings{yu2022metaformer, title={Metaformer is actually what you need for vision}, author={Yu, Weihao and Luo, Mi and Zhou, Pan and Si, Chenyang and Zhou, Yichen and Wang, Xinchao and Feng, Jiashi and Yan, Shuicheng}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition}, pages={10819--10829}, year={2022} } ```

timm/convmixer_768_32.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: mit datasets: - imagenet-1k --- # Model card for convmixer_768_32.in1k A ConvMixer image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 21.1 - GMACs: 19.5 - Activations (M): 26.0 - Image size: 224 x 224 - **Papers:** - Patches Are All You Need?: https://arxiv.org/abs/2201.09792 - **Dataset:** ImageNet-1k - **Original:** https://github.com/locuslab/convmixer ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('convmixer_768_32.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'convmixer_768_32.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 768, 32, 32) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{Chen2021CrossViTCM, title={CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification}, author={Chun-Fu Chen and Quanfu Fan and Rameswar Panda}, journal={2021 IEEE/CVF International Conference on Computer Vision (ICCV)}, year={2021}, pages={347-356} } ```

Qwen/Qwen1.5-0.5B-Chat

Qwen

transformers

text-generation

--- license: other license_name: tongyi-qianwen-research license_link: >- https://huggingface.co/Qwen/Qwen1.5-0.5B-Chat/blob/main/LICENSE language: - en pipeline_tag: text-generation tags: - chat --- # Qwen1.5-0.5B-Chat ## Introduction Qwen1.5 is the beta version of Qwen2, a transformer-based decoder-only language model pretrained on a large amount of data. In comparison with the previous released Qwen, the improvements include: * 8 model sizes, including 0.5B, 1.8B, 4B, 7B, 14B, 32B and 72B dense models, and an MoE model of 14B with 2.7B activated; * Significant performance improvement in human preference for chat models; * Multilingual support of both base and chat models; * Stable support of 32K context length for models of all sizes * No need of `trust_remote_code`. For more details, please refer to our [blog post](https://qwenlm.github.io/blog/qwen1.5/) and [GitHub repo](https://github.com/QwenLM/Qwen1.5). <br> ## Model Details Qwen1.5 is a language model series including decoder language models of different model sizes. For each size, we release the base language model and the aligned chat model. It is based on the Transformer architecture with SwiGLU activation, attention QKV bias, group query attention, mixture of sliding window attention and full attention, etc. Additionally, we have an improved tokenizer adaptive to multiple natural languages and codes. For the beta version, temporarily we did not include GQA (except for 32B) and the mixture of SWA and full attention. ## Training details We pretrained the models with a large amount of data, and we post-trained the models with both supervised finetuning and direct preference optimization. ## Requirements The code of Qwen1.5 has been in the latest Hugging face transformers and we advise you to install `transformers>=4.37.0`, or you might encounter the following error: ``` KeyError: 'qwen2' ``` ## Quickstart Here provides a code snippet with `apply_chat_template` to show you how to load the tokenizer and model and how to generate contents. ```python from transformers import AutoModelForCausalLM, AutoTokenizer device = "cuda" # the device to load the model onto model = AutoModelForCausalLM.from_pretrained( "Qwen/Qwen1.5-0.5B-Chat", torch_dtype="auto", device_map="auto" ) tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen1.5-0.5B-Chat") prompt = "Give me a short introduction to large language model." messages = [ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": prompt} ] text = tokenizer.apply_chat_template( messages, tokenize=False, add_generation_prompt=True ) model_inputs = tokenizer([text], return_tensors="pt").to(device) generated_ids = model.generate( model_inputs.input_ids, max_new_tokens=512 ) generated_ids = [ output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids) ] response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0] ``` For quantized models, we advise you to use the GPTQ, AWQ, and GGUF correspondents, namely `Qwen1.5-0.5B-Chat-GPTQ-Int4`, `Qwen1.5-0.5B-Chat-GPTQ-Int8`, `Qwen1.5-0.5B-Chat-AWQ`, and `Qwen1.5-0.5B-Chat-GGUF`. ## Tips * If you encounter code switching or other bad cases, we advise you to use our provided hyper-parameters in `generation_config.json`. ## Citation If you find our work helpful, feel free to give us a cite. ``` @article{qwen, title={Qwen Technical Report}, author={Jinze Bai and Shuai Bai and Yunfei Chu and Zeyu Cui and Kai Dang and Xiaodong Deng and Yang Fan and Wenbin Ge and Yu Han and Fei Huang and Binyuan Hui and Luo Ji and Mei Li and Junyang Lin and Runji Lin and Dayiheng Liu and Gao Liu and Chengqiang Lu and Keming Lu and Jianxin Ma and Rui Men and Xingzhang Ren and Xuancheng Ren and Chuanqi Tan and Sinan Tan and Jianhong Tu and Peng Wang and Shijie Wang and Wei Wang and Shengguang Wu and Benfeng Xu and Jin Xu and An Yang and Hao Yang and Jian Yang and Shusheng Yang and Yang Yao and Bowen Yu and Hongyi Yuan and Zheng Yuan and Jianwei Zhang and Xingxuan Zhang and Yichang Zhang and Zhenru Zhang and Chang Zhou and Jingren Zhou and Xiaohuan Zhou and Tianhang Zhu}, journal={arXiv preprint arXiv:2309.16609}, year={2023} } ```

timm/res2net50_14w_8s.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: unknown datasets: - imagenet-1k --- # Model card for res2net50_14w_8s.in1k A Res2Net (Multi-Scale ResNet) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 25.1 - GMACs: 4.2 - Activations (M): 13.3 - Image size: 224 x 224 - **Papers:** - Res2Net: A New Multi-scale Backbone Architecture: https://arxiv.org/abs/1904.01169 - **Dataset:** ImageNet-1k - **Original:** https://github.com/gasvn/Res2Net/ ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('res2net50_14w_8s.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'res2net50_14w_8s.in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 112, 112]) # torch.Size([1, 256, 56, 56]) # torch.Size([1, 512, 28, 28]) # torch.Size([1, 1024, 14, 14]) # torch.Size([1, 2048, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'res2net50_14w_8s.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{gao2019res2net, title={Res2Net: A New Multi-scale Backbone Architecture}, author={Gao, Shang-Hua and Cheng, Ming-Ming and Zhao, Kai and Zhang, Xin-Yu and Yang, Ming-Hsuan and Torr, Philip}, journal={IEEE TPAMI}, doi={10.1109/TPAMI.2019.2938758}, } ```

timm/dpn107.mx_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for dpn107.mx_in1k A DPN (Dual-Path Net) image classification model. Trained on ImageNet-1k in MXNet by paper authors and ported to PyTorch by Ross Wightman. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 86.9 - GMACs: 18.4 - Activations (M): 33.5 - Image size: 224 x 224 - **Papers:** - Dual Path Networks: https://arxiv.org/abs/1707.01629 - **Dataset:** ImageNet-1k - **Original:** https://github.com/cypw/DPNs ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('dpn107.mx_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'dpn107.mx_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 128, 112, 112]) # torch.Size([1, 376, 56, 56]) # torch.Size([1, 1152, 28, 28]) # torch.Size([1, 2432, 14, 14]) # torch.Size([1, 2688, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'dpn107.mx_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2688, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Citation ```bibtex @article{Chen2017, title={Dual Path Networks}, author={Yunpeng Chen, Jianan Li, Huaxin Xiao, Xiaojie Jin, Shuicheng Yan, Jiashi Feng}, journal={arXiv preprint arXiv:1707.01629}, year={2017} } ```

timm/repvgg_a2.rvgg_in1k

timm

timm

image-classification

--- license: mit library_name: timm tags: - image-classification - timm datasets: - imagenet-1k --- # Model card for repvgg_a2.rvgg_in1k A RepVGG image classification model. Trained on ImageNet-1k by paper authors. This model architecture is implemented using `timm`'s flexible [BYOBNet (Bring-Your-Own-Blocks Network)](https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/byobnet.py). BYOBNet allows configuration of: * block / stage layout * stem layout * output stride (dilation) * activation and norm layers * channel and spatial / self-attention layers ...and also includes `timm` features common to many other architectures, including: * stochastic depth * gradient checkpointing * layer-wise LR decay * per-stage feature extraction ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 28.2 - GMACs: 5.7 - Activations (M): 6.3 - Image size: 224 x 224 - **Papers:** - RepVGG: Making VGG-style ConvNets Great Again: https://arxiv.org/abs/2101.03697 - **Dataset:** ImageNet-1k - **Original:** https://github.com/DingXiaoH/RepVGG ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('repvgg_a2.rvgg_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'repvgg_a2.rvgg_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 112, 112]) # torch.Size([1, 96, 56, 56]) # torch.Size([1, 192, 28, 28]) # torch.Size([1, 384, 14, 14]) # torch.Size([1, 1408, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'repvgg_a2.rvgg_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1408, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ``` ```bibtex @inproceedings{ding2021repvgg, title={Repvgg: Making vgg-style convnets great again}, author={Ding, Xiaohan and Zhang, Xiangyu and Ma, Ningning and Han, Jungong and Ding, Guiguang and Sun, Jian}, booktitle={Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition}, pages={13733--13742}, year={2021} } ```

Systran/faster-whisper-base.en

Systran

ctranslate2

automatic-speech-recognition

--- language: - en tags: - audio - automatic-speech-recognition license: mit library_name: ctranslate2 --- # Whisper base.en model for CTranslate2 This repository contains the conversion of [openai/whisper-base.en](https://huggingface.co/openai/whisper-base.en) to the [CTranslate2](https://github.com/OpenNMT/CTranslate2) model format. This model can be used in CTranslate2 or projects based on CTranslate2 such as [faster-whisper](https://github.com/systran/faster-whisper). ## Example ```python from faster_whisper import WhisperModel model = WhisperModel("base.en") segments, info = model.transcribe("audio.mp3") for segment in segments: print("[%.2fs -> %.2fs] %s" % (segment.start, segment.end, segment.text)) ``` ## Conversion details The original model was converted with the following command: ``` ct2-transformers-converter --model openai/whisper-base.en --output_dir faster-whisper-base.en \ --copy_files tokenizer.json --quantization float16 ``` Note that the model weights are saved in FP16. This type can be changed when the model is loaded using the [`compute_type` option in CTranslate2](https://opennmt.net/CTranslate2/quantization.html). ## More information **For more information about the original model, see its [model card](https://huggingface.co/openai/whisper-base.en).**

RichardErkhov/davidkim205_-_Rhea-72b-v0.5-gguf

RichardErkhov

Entry not found

wangfuyun/AnimateLCM

wangfuyun

diffusers

text-to-video

--- pipeline_tag: text-to-video --- # AnimateLCM for Fast Video Generation in 4 steps. [AnimateLCM: Accelerating the Animation of Personalized Diffusion Models and Adapters with Decoupled Consistency Learning](https://arxiv.org/abs/2402.00769) by Fu-Yun Wang et al. ## We also support fast image-to-video generation, please see [AnimateLCM-SVD-xt](https://huggingface.co/wangfuyun/AnimateLCM-SVD-xt) and [AnimateLCM-I2V](https://huggingface.co/wangfuyun/AnimateLCM-I2V). For more details, please refer to our [[paper](https://arxiv.org/abs/2402.00769)] | [[code](https://github.com/G-U-N/AnimateLCM)] | [[proj-page](https://animatelcm.github.io/)] | [[civitai](https://civitai.com/models/290375/animatelcm-fast-video-generation)]. <video controls autoplay src="https://cdn-uploads.huggingface.co/production/uploads/63e9e92f20c109718713f5eb/KCwSoZCdxkkmtDg1LuXsP.mp4"></video> ## Using AnimateLCM with Diffusers ```python import torch from diffusers import AnimateDiffPipeline, LCMScheduler, MotionAdapter from diffusers.utils import export_to_gif adapter = MotionAdapter.from_pretrained("wangfuyun/AnimateLCM", torch_dtype=torch.float16) pipe = AnimateDiffPipeline.from_pretrained("emilianJR/epiCRealism", motion_adapter=adapter, torch_dtype=torch.float16) pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config, beta_schedule="linear") pipe.load_lora_weights("wangfuyun/AnimateLCM", weight_name="AnimateLCM_sd15_t2v_lora.safetensors", adapter_name="lcm-lora") pipe.set_adapters(["lcm-lora"], [0.8]) pipe.enable_vae_slicing() pipe.enable_model_cpu_offload() output = pipe( prompt="A space rocket with trails of smoke behind it launching into space from the desert, 4k, high resolution", negative_prompt="bad quality, worse quality, low resolution", num_frames=16, guidance_scale=2.0, num_inference_steps=6, generator=torch.Generator("cpu").manual_seed(0), ) frames = output.frames[0] export_to_gif(frames, "animatelcm.gif") ```

mradermacher/AceGPT-v2-32B-GGUF

mradermacher

transformers

--- base_model: FreedomIntelligence/AceGPT-v2-32B language: - ar - zh - en library_name: transformers license: apache-2.0 quantized_by: mradermacher --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: --> static quants of https://huggingface.co/FreedomIntelligence/AceGPT-v2-32B <!-- provided-files --> weighted/imatrix quants are available at https://huggingface.co/mradermacher/AceGPT-v2-32B-i1-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q2_K.gguf) | Q2_K | 12.3 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.IQ3_XS.gguf) | IQ3_XS | 13.7 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q3_K_S.gguf) | Q3_K_S | 14.4 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.IQ3_S.gguf) | IQ3_S | 14.4 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.IQ3_M.gguf) | IQ3_M | 14.8 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q3_K_M.gguf) | Q3_K_M | 15.9 | lower quality | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q3_K_L.gguf) | Q3_K_L | 17.2 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.IQ4_XS.gguf) | IQ4_XS | 17.8 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q4_K_S.gguf) | Q4_K_S | 18.7 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q4_K_M.gguf) | Q4_K_M | 19.8 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q5_K_S.gguf) | Q5_K_S | 22.6 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q5_K_M.gguf) | Q5_K_M | 23.2 | | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q6_K.gguf) | Q6_K | 26.8 | very good quality | | [GGUF](https://huggingface.co/mradermacher/AceGPT-v2-32B-GGUF/resolve/main/AceGPT-v2-32B.Q8_0.gguf) | Q8_0 | 34.7 | fast, best quality | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):  And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. <!-- end -->

ptx0/terminus-xl-refiner

ptx0

diffusers

text-to-image

--- license: creativeml-openrail-m base_model: "segmind/SSD-1B" tags: - stable-diffusion - stable-diffusion-diffusers - text-to-image - diffusers - full inference: true --- # terminus-xl-refiner This is a full rank finetune derived from [segmind/SSD-1B](https://huggingface.co/segmind/SSD-1B). The main validation prompt used during training was: ``` a cute anime character named toast ``` ## Validation settings - CFG: `7.5` - CFG Rescale: `0.7` - Steps: `30` - Sampler: `ddpm` - Seed: `420420420` - Resolution: `1024` Note: The validation settings are not necessarily the same as the [training settings](#training-settings). <Gallery /> The text encoder **was not** trained. You may reuse the base model text encoder for inference. ## Training settings - Training epochs: 0 - Training steps: 12800 - Learning rate: 2e-06 - Effective batch size: 16 - Micro-batch size: 4 - Gradient accumulation steps: 4 - Number of GPUs: 1 - Prediction type: v_prediction - Rescaled betas zero SNR: True - Optimizer: AdamW, stochastic bf16 - Precision: Pure BF16 - Xformers: Enabled ## Datasets ### pixel-art - Repeats: 0 - Total number of images: 1040 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### signs - Repeats: 0 - Total number of images: 368 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### experimental - Repeats: 0 - Total number of images: 3024 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### ethnic - Repeats: 0 - Total number of images: 3072 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### sports - Repeats: 0 - Total number of images: 784 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### architecture - Repeats: 0 - Total number of images: 4336 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### shutterstock - Repeats: 0 - Total number of images: 21072 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### cinemamix-1mp - Repeats: 0 - Total number of images: 9008 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### nsfw-1024 - Repeats: 0 - Total number of images: 10800 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### anatomy - Repeats: 5 - Total number of images: 16417 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### yoga - Repeats: 0 - Total number of images: 3600 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### photo-aesthetics - Repeats: 0 - Total number of images: 33136 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### text-1mp - Repeats: 5 - Total number of images: 13170 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ### photo-concept-bucket - Repeats: 0 - Total number of images: 567554 - Total number of aspect buckets: 3 - Resolution: 1.0 megapixels - Cropped: True - Crop style: random - Crop aspect: random ## Inference ```python import torch from diffusers import DiffusionPipeline model_id = "terminus-xl-refiner" prompt = "a cute anime character named toast" negative_prompt = "malformed, disgusting, overexposed, washed-out" pipeline = DiffusionPipeline.from_pretrained(model_id) pipeline.to('cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu') image = pipeline( prompt=prompt, negative_prompt='blurry, cropped, ugly', num_inference_steps=30, generator=torch.Generator(device='cuda' if torch.cuda.is_available() else 'mps' if torch.backends.mps.is_available() else 'cpu').manual_seed(1641421826), width=1152, height=768, guidance_scale=7.5, guidance_rescale=0.7, ).images[0] image.save("output.png", format="PNG") ```

dataautogpt3/ProteusV0.2

dataautogpt3

diffusers

text-to-image

--- pipeline_tag: text-to-image widget: - text: >- black fluffy gorgeous dangerous cat animal creature, large orange eyes, big fluffy ears, piercing gaze, full moon, dark ambiance, best quality, extremely detailed output: url: ComfyUI_03087_.png - text: >- (impressionistic realism by csybgh), a 50 something male, working in banking, very short dyed dark curly balding hair, Afro-Asiatic ancestry, talks a lot but listens poorly, stuck in the past, wearing a suit, he has a certain charm, bronze skintone, sitting in a bar at night, he is smoking and feeling cool, drunk on plum wine, masterpiece, 8k, hyper detailed, smokey ambiance, perfect hands AND fingers output: url: GEN8-iTXcAA-okN.jpeg - text: >- high quality pixel art, a pixel art silhouette of an anime space-themed girl in a space-punk steampunk style, lying in her bed by the window of a spaceship, smoking, with a rustic feel. The image should embody epic portraiture and double exposure, featuring an isolated landscape visible through the window. The colors should primarily be dynamic and action-packed, with a strong use of negative space. The entire artwork should be in pixel art style, emphasizing the characters shape and set against a white background. Silhouette output: url: ComfyUI_03060_.png - text: >- The image features an older man, a long white beard and mustache, He has a stern expression, giving the impression of a wise and experienced individual. The mans beard and mustache are prominent, adding to his distinguished appearance. The close-up shot of the mans face emphasizes his facial features and the intensity of his gaze. output: url: ComfyUI_03017_.png - text: >- Super Closeup Portrait, action shot, Profoundly dark whitish meadow, glass flowers, Stains, space grunge style, Jeanne d'Arc wearing White Olive green used styled Cotton frock, Wielding thin silver sword, Sci-fi vibe, dirty, noisy, Vintage monk style, very detailed, hd output: url: ComfyUI_03045.png - text: >- cinematic film still of Kodak Motion Picture Film: (Sharp Detailed Image) An Oscar winning movie for Best Cinematography a woman in a kimono standing on a subway train in Japan Kodak Motion Picture Film Style, shallow depth of field, vignette, highly detailed, high budget, bokeh, cinemascope, moody, epic, gorgeous, film grain, grainy output: url: 3.png - text: >- in the style of artgerm, comic style,3D model, mythical seascape, negative space, space quixotic dreams, temporal hallucination, psychedelic, mystical, intricate details, very bright neon colors, (vantablack background:1.5), pointillism, pareidolia, melting, symbolism, very high contrast, chiaroscuro parameters: negative_prompt: >- bad quality, bad anatomy, worst quality, low quality, low resolutions, extra fingers, blur, blurry, ugly, wrongs proportions, watermark, image artifacts, lowres, ugly, jpeg artifacts, deformed, noisy image output: url: ComfyUI_03061_.png - text: >- 1980s anime portrait of a character glitching. His face is separated from his body by heavy static. His face is deformed by pain. Dream-like, analog horror, glitch, terrifying output: url: ComfyUI_03092_.png - text: (("Proteus"):text_logo:1) output: url: ComfyUI_03297_.png - text: >- dan seagrave, dante, Abandon All Hope, Ye Who Enter Here, hell religious art purgatory zdzislaw Beksinski, abyss inferno, lost, wanderer output: url: ComfyUI_03483_.png license: gpl-3.0 --- <Gallery /> ## ProteusV0.2 merged with RealCartoonXL to fix issues with inability to understand tags related to anime or cartoon styles at just a weight of 0.5% out of 100% using custom scripts with slerp like methods. Version 0.2 shows subtle yet significant improvements over Version 0.1. It demonstrates enhanced prompt understanding that surpasses MJ6, while also approaching its stylistic capabilities. ## Proteus Proteus serves as a sophisticated enhancement over OpenDalleV1.1, leveraging its core functionalities to deliver superior outcomes. Key areas of advancement include heightened responsiveness to prompts and augmented creative capacities. To achieve this, it was fine-tuned using approximately 220,000 GPTV captioned images from copyright-free stock images (with some anime included), which were then normalized. Additionally, DPO (Direct Preference Optimization) was employed through a collection of 10,000 carefully selected high-quality, AI-generated image pairs. In pursuit of optimal performance, numerous LORA (Low-Rank Adaptation) models are trained independently before being selectively incorporated into the principal model via dynamic application methods. These techniques involve targeting particular segments within the model while avoiding interference with other areas during the learning phase. Consequently, Proteus exhibits marked improvements in portraying intricate facial characteristics and lifelike skin textures, all while sustaining commendable proficiency across various aesthetic domains, notably surrealism, anime, and cartoon-style visualizations. ## Settings for ProteusV0.2 Use these settings for the best results with ProteusV0.2: CFG Scale: Use a CFG scale of 8 to 7 Steps: 20 to 60 steps for more detail, 20 steps for faster results. Sampler: DPM++ 2M SDE Scheduler: Karras Resolution: 1280x1280 or 1024x1024 please also consider using these keep words to improve your prompts: best quality, HD, `~*~aesthetic~*~`. if you are having trouble coming up with prompts you can use this GPT I put together to help you refine the prompt. https://chat.openai.com/g/g-RziQNoydR-diffusion-master ## Use it with 🧨 diffusers ```python import torch from diffusers import ( StableDiffusionXLPipeline, KDPM2AncestralDiscreteScheduler, AutoencoderKL ) # Load VAE component vae = AutoencoderKL.from_pretrained( "madebyollin/sdxl-vae-fp16-fix", torch_dtype=torch.float16 ) # Configure the pipeline pipe = StableDiffusionXLPipeline.from_pretrained( "dataautogpt3/ProteusV0.2", vae=vae, torch_dtype=torch.float16 ) pipe.scheduler = KDPM2AncestralDiscreteScheduler.from_config(pipe.scheduler.config) pipe.to('cuda') # Define prompts and generate image prompt = "black fluffy gorgeous dangerous cat animal creature, large orange eyes, big fluffy ears, piercing gaze, full moon, dark ambiance, best quality, extremely detailed" negative_prompt = "nsfw, bad quality, bad anatomy, worst quality, low quality, low resolutions, extra fingers, blur, blurry, ugly, wrongs proportions, watermark, image artifacts, lowres, ugly, jpeg artifacts, deformed, noisy image" image = pipe( prompt, negative_prompt=negative_prompt, width=1024, height=1024, guidance_scale=7.5, num_inference_steps=50 ).images[0] ``` please support the work I do through donating to me on: https://www.buymeacoffee.com/DataVoid or following me on https://twitter.com/DataPlusEngine

timm/visformer_small.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for visformer_small.in1k A Visformer image classification model. Trained on ImageNet-1k by https://github.com/hzhang57 and https://github.com/developer0hye. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 40.2 - GMACs: 4.9 - Activations (M): 11.4 - Image size: 224 x 224 - **Papers:** - Visformer: The Vision-friendly Transformer: https://arxiv.org/abs/2104.12533 - **Dataset:** ImageNet-1k - **Original:** https://github.com/danczs/Visformer ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('visformer_small.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'visformer_small.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 768, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @inproceedings{chen2021visformer, title={Visformer: The vision-friendly transformer}, author={Chen, Zhengsu and Xie, Lingxi and Niu, Jianwei and Liu, Xuefeng and Wei, Longhui and Tian, Qi}, booktitle={Proceedings of the IEEE/CVF international conference on computer vision}, pages={589--598}, year={2021} } ```

timm/inception_v3.gluon_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for inception_v3.gluon_in1k A Inception-v3 image classification model. Trained on ImageNet-1k by MxNet GLUON authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 23.8 - GMACs: 5.7 - Activations (M): 9.0 - Image size: 299 x 299 - **Papers:** - Rethinking the Inception Architecture for Computer Vision: https://arxiv.org/abs/1512.00567 - **Original:** https://github.com/tensorflow/models - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('inception_v3.gluon_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'inception_v3.gluon_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 147, 147]) # torch.Size([1, 192, 71, 71]) # torch.Size([1, 288, 35, 35]) # torch.Size([1, 768, 17, 17]) # torch.Size([1, 2048, 8, 8]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'inception_v3.gluon_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 8, 8) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @article{DBLP:journals/corr/SzegedyVISW15, author = {Christian Szegedy and Vincent Vanhoucke and Sergey Ioffe and Jonathon Shlens and Zbigniew Wojna}, title = {Rethinking the Inception Architecture for Computer Vision}, journal = {CoRR}, volume = {abs/1512.00567}, year = {2015}, url = {http://arxiv.org/abs/1512.00567}, archivePrefix = {arXiv}, eprint = {1512.00567}, timestamp = {Mon, 13 Aug 2018 16:49:07 +0200}, biburl = {https://dblp.org/rec/journals/corr/SzegedyVISW15.bib}, bibsource = {dblp computer science bibliography, https://dblp.org} } ```

TheBloke/Llama-2-7B-fp16

TheBloke

transformers

text-generation

Entry not found

llava-hf/llava-1.5-13b-hf

llava-hf

transformers

image-to-text

--- language: - en pipeline_tag: image-to-text inference: false arxiv: 2304.08485 --- # LLaVA Model Card  Below is the model card of Llava model 13b, which is copied from the original Llava model card that you can find [here](https://huggingface.co/liuhaotian/llava-v1.5-13b). Check out also the Google Colab demo to run Llava on a free-tier Google Colab instance: [](https://colab.research.google.com/drive/1qsl6cd2c8gGtEW1xV5io7S8NHh-Cp1TV?usp=sharing) Or check out our Spaces demo! [](https://huggingface.co/spaces/llava-hf/llava-4bit) ## Model details **Model type:** LLaVA is an open-source chatbot trained by fine-tuning LLaMA/Vicuna on GPT-generated multimodal instruction-following data. It is an auto-regressive language model, based on the transformer architecture. **Model date:** LLaVA-v1.5-13B was trained in September 2023. **Paper or resources for more information:** https://llava-vl.github.io/ ## How to use the model First, make sure to have `transformers >= 4.35.3`. The model supports multi-image and multi-prompt generation. Meaning that you can pass multiple images in your prompt. Make sure also to follow the correct prompt template (`USER: xxx\nASSISTANT:`) and add the token `<image>` to the location where you want to query images: ### Using `pipeline`: Below we used [`"llava-hf/llava-1.5-13b-hf"`](https://huggingface.co/llava-hf/llava-1.5-13b-hf) checkpoint. ```python from transformers import pipeline from PIL import Image import requests model_id = "llava-hf/llava-1.5-13b-hf" pipe = pipeline("image-to-text", model=model_id) url = "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/transformers/tasks/ai2d-demo.jpg" image = Image.open(requests.get(url, stream=True).raw) prompt = "USER: <image>\nWhat does the label 15 represent? (1) lava (2) core (3) tunnel (4) ash cloud\nASSISTANT:" outputs = pipe(image, prompt=prompt, generate_kwargs={"max_new_tokens": 200}) print(outputs) >>> {"generated_text": "\nUSER: What does the label 15 represent? (1) lava (2) core (3) tunnel (4) ash cloud\nASSISTANT: Lava"} ``` ### Using pure `transformers`: Below is an example script to run generation in `float16` precision on a GPU device: ```python import requests from PIL import Image import torch from transformers import AutoProcessor, LlavaForConditionalGeneration model_id = "llava-hf/llava-1.5-13b-hf" prompt = "USER: <image>\nWhat are these?\nASSISTANT:" image_file = "http://images.cocodataset.org/val2017/000000039769.jpg" model = LlavaForConditionalGeneration.from_pretrained( model_id, torch_dtype=torch.float16, low_cpu_mem_usage=True, ).to(0) processor = AutoProcessor.from_pretrained(model_id) raw_image = Image.open(requests.get(image_file, stream=True).raw) inputs = processor(prompt, raw_image, return_tensors='pt').to(0, torch.float16) output = model.generate(**inputs, max_new_tokens=200, do_sample=False) print(processor.decode(output[0][2:], skip_special_tokens=True)) ``` ### Model optimization #### 4-bit quantization through `bitsandbytes` library First make sure to install `bitsandbytes`, `pip install bitsandbytes` and make sure to have access to a CUDA compatible GPU device. Simply change the snippet above with: ```diff model = LlavaForConditionalGeneration.from_pretrained( model_id, torch_dtype=torch.float16, low_cpu_mem_usage=True, + load_in_4bit=True ) ``` #### Use Flash-Attention 2 to further speed-up generation First make sure to install `flash-attn`. Refer to the [original repository of Flash Attention](https://github.com/Dao-AILab/flash-attention) regarding that package installation. Simply change the snippet above with: ```diff model = LlavaForConditionalGeneration.from_pretrained( model_id, torch_dtype=torch.float16, low_cpu_mem_usage=True, + use_flash_attention_2=True ).to(0) ``` ## License Llama 2 is licensed under the LLAMA 2 Community License, Copyright (c) Meta Platforms, Inc. All Rights Reserved.

timm/regnety_032.ra_in1k

timm

timm

image-classification

--- license: apache-2.0 library_name: timm tags: - image-classification - timm datasets: - imagenet-1k --- # Model card for regnety_032.ra_in1k A RegNetY-3.2GF image classification model. Trained on ImageNet-1k by Ross Wightman in `timm`. The `timm` RegNet implementation includes a number of enhancements not present in other implementations, including: * stochastic depth * gradient checkpointing * layer-wise LR decay * configurable output stride (dilation) * configurable activation and norm layers * option for a pre-activation bottleneck block used in RegNetV variant * only known RegNetZ model definitions with pretrained weights ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 19.4 - GMACs: 3.2 - Activations (M): 11.3 - Image size: train = 224 x 224, test = 288 x 288 - **Papers:** - Designing Network Design Spaces: https://arxiv.org/abs/2003.13678 - **Dataset:** ImageNet-1k - **Original:** https://github.com/huggingface/pytorch-image-models ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('regnety_032.ra_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'regnety_032.ra_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 32, 112, 112]) # torch.Size([1, 72, 56, 56]) # torch.Size([1, 216, 28, 28]) # torch.Size([1, 576, 14, 14]) # torch.Size([1, 1512, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'regnety_032.ra_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1512, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). For the comparison summary below, the ra_in1k, ra3_in1k, ch_in1k, sw_*, and lion_* tagged weights are trained in `timm`. |model |img_size|top1 |top5 |param_count|gmacs|macts | |-------------------------|--------|------|------|-----------|-----|------| |[regnety_1280.swag_ft_in1k](https://huggingface.co/timm/regnety_1280.swag_ft_in1k)|384 |88.228|98.684|644.81 |374.99|210.2 | |[regnety_320.swag_ft_in1k](https://huggingface.co/timm/regnety_320.swag_ft_in1k)|384 |86.84 |98.364|145.05 |95.0 |88.87 | |[regnety_160.swag_ft_in1k](https://huggingface.co/timm/regnety_160.swag_ft_in1k)|384 |86.024|98.05 |83.59 |46.87|67.67 | |[regnety_160.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.sw_in12k_ft_in1k)|288 |86.004|97.83 |83.59 |26.37|38.07 | |[regnety_1280.swag_lc_in1k](https://huggingface.co/timm/regnety_1280.swag_lc_in1k)|224 |85.996|97.848|644.81 |127.66|71.58 | |[regnety_160.lion_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.lion_in12k_ft_in1k)|288 |85.982|97.844|83.59 |26.37|38.07 | |[regnety_160.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.sw_in12k_ft_in1k)|224 |85.574|97.666|83.59 |15.96|23.04 | |[regnety_160.lion_in12k_ft_in1k](https://huggingface.co/timm/regnety_160.lion_in12k_ft_in1k)|224 |85.564|97.674|83.59 |15.96|23.04 | |[regnety_120.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_120.sw_in12k_ft_in1k)|288 |85.398|97.584|51.82 |20.06|35.34 | |[regnety_2560.seer_ft_in1k](https://huggingface.co/timm/regnety_2560.seer_ft_in1k)|384 |85.15 |97.436|1282.6 |747.83|296.49| |[regnetz_e8.ra3_in1k](https://huggingface.co/timm/regnetz_e8.ra3_in1k)|320 |85.036|97.268|57.7 |15.46|63.94 | |[regnety_120.sw_in12k_ft_in1k](https://huggingface.co/timm/regnety_120.sw_in12k_ft_in1k)|224 |84.976|97.416|51.82 |12.14|21.38 | |[regnety_320.swag_lc_in1k](https://huggingface.co/timm/regnety_320.swag_lc_in1k)|224 |84.56 |97.446|145.05 |32.34|30.26 | |[regnetz_040_h.ra3_in1k](https://huggingface.co/timm/regnetz_040_h.ra3_in1k)|320 |84.496|97.004|28.94 |6.43 |37.94 | |[regnetz_e8.ra3_in1k](https://huggingface.co/timm/regnetz_e8.ra3_in1k)|256 |84.436|97.02 |57.7 |9.91 |40.94 | |[regnety_1280.seer_ft_in1k](https://huggingface.co/timm/regnety_1280.seer_ft_in1k)|384 |84.432|97.092|644.81 |374.99|210.2 | |[regnetz_040.ra3_in1k](https://huggingface.co/timm/regnetz_040.ra3_in1k)|320 |84.246|96.93 |27.12 |6.35 |37.78 | |[regnetz_d8.ra3_in1k](https://huggingface.co/timm/regnetz_d8.ra3_in1k)|320 |84.054|96.992|23.37 |6.19 |37.08 | |[regnetz_d8_evos.ch_in1k](https://huggingface.co/timm/regnetz_d8_evos.ch_in1k)|320 |84.038|96.992|23.46 |7.03 |38.92 | |[regnetz_d32.ra3_in1k](https://huggingface.co/timm/regnetz_d32.ra3_in1k)|320 |84.022|96.866|27.58 |9.33 |37.08 | |[regnety_080.ra3_in1k](https://huggingface.co/timm/regnety_080.ra3_in1k)|288 |83.932|96.888|39.18 |13.22|29.69 | |[regnety_640.seer_ft_in1k](https://huggingface.co/timm/regnety_640.seer_ft_in1k)|384 |83.912|96.924|281.38 |188.47|124.83| |[regnety_160.swag_lc_in1k](https://huggingface.co/timm/regnety_160.swag_lc_in1k)|224 |83.778|97.286|83.59 |15.96|23.04 | |[regnetz_040_h.ra3_in1k](https://huggingface.co/timm/regnetz_040_h.ra3_in1k)|256 |83.776|96.704|28.94 |4.12 |24.29 | |[regnetv_064.ra3_in1k](https://huggingface.co/timm/regnetv_064.ra3_in1k)|288 |83.72 |96.75 |30.58 |10.55|27.11 | |[regnety_064.ra3_in1k](https://huggingface.co/timm/regnety_064.ra3_in1k)|288 |83.718|96.724|30.58 |10.56|27.11 | |[regnety_160.deit_in1k](https://huggingface.co/timm/regnety_160.deit_in1k)|288 |83.69 |96.778|83.59 |26.37|38.07 | |[regnetz_040.ra3_in1k](https://huggingface.co/timm/regnetz_040.ra3_in1k)|256 |83.62 |96.704|27.12 |4.06 |24.19 | |[regnetz_d8.ra3_in1k](https://huggingface.co/timm/regnetz_d8.ra3_in1k)|256 |83.438|96.776|23.37 |3.97 |23.74 | |[regnetz_d32.ra3_in1k](https://huggingface.co/timm/regnetz_d32.ra3_in1k)|256 |83.424|96.632|27.58 |5.98 |23.74 | |[regnetz_d8_evos.ch_in1k](https://huggingface.co/timm/regnetz_d8_evos.ch_in1k)|256 |83.36 |96.636|23.46 |4.5 |24.92 | |[regnety_320.seer_ft_in1k](https://huggingface.co/timm/regnety_320.seer_ft_in1k)|384 |83.35 |96.71 |145.05 |95.0 |88.87 | |[regnetv_040.ra3_in1k](https://huggingface.co/timm/regnetv_040.ra3_in1k)|288 |83.204|96.66 |20.64 |6.6 |20.3 | |[regnety_320.tv2_in1k](https://huggingface.co/timm/regnety_320.tv2_in1k)|224 |83.162|96.42 |145.05 |32.34|30.26 | |[regnety_080.ra3_in1k](https://huggingface.co/timm/regnety_080.ra3_in1k)|224 |83.16 |96.486|39.18 |8.0 |17.97 | |[regnetv_064.ra3_in1k](https://huggingface.co/timm/regnetv_064.ra3_in1k)|224 |83.108|96.458|30.58 |6.39 |16.41 | |[regnety_040.ra3_in1k](https://huggingface.co/timm/regnety_040.ra3_in1k)|288 |83.044|96.5 |20.65 |6.61 |20.3 | |[regnety_064.ra3_in1k](https://huggingface.co/timm/regnety_064.ra3_in1k)|224 |83.02 |96.292|30.58 |6.39 |16.41 | |[regnety_160.deit_in1k](https://huggingface.co/timm/regnety_160.deit_in1k)|224 |82.974|96.502|83.59 |15.96|23.04 | |[regnetx_320.tv2_in1k](https://huggingface.co/timm/regnetx_320.tv2_in1k)|224 |82.816|96.208|107.81 |31.81|36.3 | |[regnety_032.ra_in1k](https://huggingface.co/timm/regnety_032.ra_in1k)|288 |82.742|96.418|19.44 |5.29 |18.61 | |[regnety_160.tv2_in1k](https://huggingface.co/timm/regnety_160.tv2_in1k)|224 |82.634|96.22 |83.59 |15.96|23.04 | |[regnetz_c16_evos.ch_in1k](https://huggingface.co/timm/regnetz_c16_evos.ch_in1k)|320 |82.634|96.472|13.49 |3.86 |25.88 | |[regnety_080_tv.tv2_in1k](https://huggingface.co/timm/regnety_080_tv.tv2_in1k)|224 |82.592|96.246|39.38 |8.51 |19.73 | |[regnetx_160.tv2_in1k](https://huggingface.co/timm/regnetx_160.tv2_in1k)|224 |82.564|96.052|54.28 |15.99|25.52 | |[regnetz_c16.ra3_in1k](https://huggingface.co/timm/regnetz_c16.ra3_in1k)|320 |82.51 |96.358|13.46 |3.92 |25.88 | |[regnetv_040.ra3_in1k](https://huggingface.co/timm/regnetv_040.ra3_in1k)|224 |82.44 |96.198|20.64 |4.0 |12.29 | |[regnety_040.ra3_in1k](https://huggingface.co/timm/regnety_040.ra3_in1k)|224 |82.304|96.078|20.65 |4.0 |12.29 | |[regnetz_c16.ra3_in1k](https://huggingface.co/timm/regnetz_c16.ra3_in1k)|256 |82.16 |96.048|13.46 |2.51 |16.57 | |[regnetz_c16_evos.ch_in1k](https://huggingface.co/timm/regnetz_c16_evos.ch_in1k)|256 |81.936|96.15 |13.49 |2.48 |16.57 | |[regnety_032.ra_in1k](https://huggingface.co/timm/regnety_032.ra_in1k)|224 |81.924|95.988|19.44 |3.2 |11.26 | |[regnety_032.tv2_in1k](https://huggingface.co/timm/regnety_032.tv2_in1k)|224 |81.77 |95.842|19.44 |3.2 |11.26 | |[regnetx_080.tv2_in1k](https://huggingface.co/timm/regnetx_080.tv2_in1k)|224 |81.552|95.544|39.57 |8.02 |14.06 | |[regnetx_032.tv2_in1k](https://huggingface.co/timm/regnetx_032.tv2_in1k)|224 |80.924|95.27 |15.3 |3.2 |11.37 | |[regnety_320.pycls_in1k](https://huggingface.co/timm/regnety_320.pycls_in1k)|224 |80.804|95.246|145.05 |32.34|30.26 | |[regnetz_b16.ra3_in1k](https://huggingface.co/timm/regnetz_b16.ra3_in1k)|288 |80.712|95.47 |9.72 |2.39 |16.43 | |[regnety_016.tv2_in1k](https://huggingface.co/timm/regnety_016.tv2_in1k)|224 |80.66 |95.334|11.2 |1.63 |8.04 | |[regnety_120.pycls_in1k](https://huggingface.co/timm/regnety_120.pycls_in1k)|224 |80.37 |95.12 |51.82 |12.14|21.38 | |[regnety_160.pycls_in1k](https://huggingface.co/timm/regnety_160.pycls_in1k)|224 |80.288|94.964|83.59 |15.96|23.04 | |[regnetx_320.pycls_in1k](https://huggingface.co/timm/regnetx_320.pycls_in1k)|224 |80.246|95.01 |107.81 |31.81|36.3 | |[regnety_080.pycls_in1k](https://huggingface.co/timm/regnety_080.pycls_in1k)|224 |79.882|94.834|39.18 |8.0 |17.97 | |[regnetz_b16.ra3_in1k](https://huggingface.co/timm/regnetz_b16.ra3_in1k)|224 |79.872|94.974|9.72 |1.45 |9.95 | |[regnetx_160.pycls_in1k](https://huggingface.co/timm/regnetx_160.pycls_in1k)|224 |79.862|94.828|54.28 |15.99|25.52 | |[regnety_064.pycls_in1k](https://huggingface.co/timm/regnety_064.pycls_in1k)|224 |79.716|94.772|30.58 |6.39 |16.41 | |[regnetx_120.pycls_in1k](https://huggingface.co/timm/regnetx_120.pycls_in1k)|224 |79.592|94.738|46.11 |12.13|21.37 | |[regnetx_016.tv2_in1k](https://huggingface.co/timm/regnetx_016.tv2_in1k)|224 |79.44 |94.772|9.19 |1.62 |7.93 | |[regnety_040.pycls_in1k](https://huggingface.co/timm/regnety_040.pycls_in1k)|224 |79.23 |94.654|20.65 |4.0 |12.29 | |[regnetx_080.pycls_in1k](https://huggingface.co/timm/regnetx_080.pycls_in1k)|224 |79.198|94.55 |39.57 |8.02 |14.06 | |[regnetx_064.pycls_in1k](https://huggingface.co/timm/regnetx_064.pycls_in1k)|224 |79.064|94.454|26.21 |6.49 |16.37 | |[regnety_032.pycls_in1k](https://huggingface.co/timm/regnety_032.pycls_in1k)|224 |78.884|94.412|19.44 |3.2 |11.26 | |[regnety_008_tv.tv2_in1k](https://huggingface.co/timm/regnety_008_tv.tv2_in1k)|224 |78.654|94.388|6.43 |0.84 |5.42 | |[regnetx_040.pycls_in1k](https://huggingface.co/timm/regnetx_040.pycls_in1k)|224 |78.482|94.24 |22.12 |3.99 |12.2 | |[regnetx_032.pycls_in1k](https://huggingface.co/timm/regnetx_032.pycls_in1k)|224 |78.178|94.08 |15.3 |3.2 |11.37 | |[regnety_016.pycls_in1k](https://huggingface.co/timm/regnety_016.pycls_in1k)|224 |77.862|93.73 |11.2 |1.63 |8.04 | |[regnetx_008.tv2_in1k](https://huggingface.co/timm/regnetx_008.tv2_in1k)|224 |77.302|93.672|7.26 |0.81 |5.15 | |[regnetx_016.pycls_in1k](https://huggingface.co/timm/regnetx_016.pycls_in1k)|224 |76.908|93.418|9.19 |1.62 |7.93 | |[regnety_008.pycls_in1k](https://huggingface.co/timm/regnety_008.pycls_in1k)|224 |76.296|93.05 |6.26 |0.81 |5.25 | |[regnety_004.tv2_in1k](https://huggingface.co/timm/regnety_004.tv2_in1k)|224 |75.592|92.712|4.34 |0.41 |3.89 | |[regnety_006.pycls_in1k](https://huggingface.co/timm/regnety_006.pycls_in1k)|224 |75.244|92.518|6.06 |0.61 |4.33 | |[regnetx_008.pycls_in1k](https://huggingface.co/timm/regnetx_008.pycls_in1k)|224 |75.042|92.342|7.26 |0.81 |5.15 | |[regnetx_004_tv.tv2_in1k](https://huggingface.co/timm/regnetx_004_tv.tv2_in1k)|224 |74.57 |92.184|5.5 |0.42 |3.17 | |[regnety_004.pycls_in1k](https://huggingface.co/timm/regnety_004.pycls_in1k)|224 |74.018|91.764|4.34 |0.41 |3.89 | |[regnetx_006.pycls_in1k](https://huggingface.co/timm/regnetx_006.pycls_in1k)|224 |73.862|91.67 |6.2 |0.61 |3.98 | |[regnetx_004.pycls_in1k](https://huggingface.co/timm/regnetx_004.pycls_in1k)|224 |72.38 |90.832|5.16 |0.4 |3.14 | |[regnety_002.pycls_in1k](https://huggingface.co/timm/regnety_002.pycls_in1k)|224 |70.282|89.534|3.16 |0.2 |2.17 | |[regnetx_002.pycls_in1k](https://huggingface.co/timm/regnetx_002.pycls_in1k)|224 |68.752|88.556|2.68 |0.2 |2.16 | ## Citation ```bibtex @InProceedings{Radosavovic2020, title = {Designing Network Design Spaces}, author = {Ilija Radosavovic and Raj Prateek Kosaraju and Ross Girshick and Kaiming He and Piotr Doll{'a}r}, booktitle = {CVPR}, year = {2020} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

mradermacher/Yi-1.5-34B-32K-i1-GGUF

mradermacher

transformers

--- base_model: 01-ai/Yi-1.5-34B-32K language: - en library_name: transformers license: apache-2.0 quantized_by: mradermacher --- ## About <!-- ### quantize_version: 2 --> <!-- ### output_tensor_quantised: 1 --> <!-- ### convert_type: hf --> <!-- ### vocab_type: --> <!-- ### tags: nicoboss --> weighted/imatrix quants of https://huggingface.co/01-ai/Yi-1.5-34B-32K <!-- provided-files --> static quants are available at https://huggingface.co/mradermacher/Yi-1.5-34B-32K-GGUF ## Usage If you are unsure how to use GGUF files, refer to one of [TheBloke's READMEs](https://huggingface.co/TheBloke/KafkaLM-70B-German-V0.1-GGUF) for more details, including on how to concatenate multi-part files. ## Provided Quants (sorted by size, not necessarily quality. IQ-quants are often preferable over similar sized non-IQ quants) | Link | Type | Size/GB | Notes | |:-----|:-----|--------:|:------| | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ1_S.gguf) | i1-IQ1_S | 7.6 | for the desperate | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ1_M.gguf) | i1-IQ1_M | 8.3 | mostly desperate | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ2_XXS.gguf) | i1-IQ2_XXS | 9.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ2_XS.gguf) | i1-IQ2_XS | 10.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ2_S.gguf) | i1-IQ2_S | 11.0 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ2_M.gguf) | i1-IQ2_M | 11.9 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q2_K.gguf) | i1-Q2_K | 12.9 | IQ3_XXS probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ3_XXS.gguf) | i1-IQ3_XXS | 13.4 | lower quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ3_XS.gguf) | i1-IQ3_XS | 14.3 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q3_K_S.gguf) | i1-Q3_K_S | 15.1 | IQ3_XS probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ3_S.gguf) | i1-IQ3_S | 15.1 | beats Q3_K* | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ3_M.gguf) | i1-IQ3_M | 15.7 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q3_K_M.gguf) | i1-Q3_K_M | 16.8 | IQ3_S probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q3_K_L.gguf) | i1-Q3_K_L | 18.2 | IQ3_M probably better | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-IQ4_XS.gguf) | i1-IQ4_XS | 18.6 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q4_0.gguf) | i1-Q4_0 | 19.6 | fast, low quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q4_K_S.gguf) | i1-Q4_K_S | 19.7 | optimal size/speed/quality | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q4_K_M.gguf) | i1-Q4_K_M | 20.8 | fast, recommended | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q5_K_S.gguf) | i1-Q5_K_S | 23.8 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q5_K_M.gguf) | i1-Q5_K_M | 24.4 | | | [GGUF](https://huggingface.co/mradermacher/Yi-1.5-34B-32K-i1-GGUF/resolve/main/Yi-1.5-34B-32K.i1-Q6_K.gguf) | i1-Q6_K | 28.3 | practically like static Q6_K | Here is a handy graph by ikawrakow comparing some lower-quality quant types (lower is better):  And here are Artefact2's thoughts on the matter: https://gist.github.com/Artefact2/b5f810600771265fc1e39442288e8ec9 ## FAQ / Model Request See https://huggingface.co/mradermacher/model_requests for some answers to questions you might have and/or if you want some other model quantized. ## Thanks I thank my company, [nethype GmbH](https://www.nethype.de/), for letting me use its servers and providing upgrades to my workstation to enable this work in my free time. Additional thanks to [@nicoboss](https://huggingface.co/nicoboss) for giving me access to his hardware for calculating the imatrix for these quants. <!-- end -->

timm/botnet26t_256.c1_in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for botnet26t_256.c1_in1k A BotNet image classification model (, based on ResNet architecture). Trained on ImageNet-1k in `timm` by Ross Wightman. NOTE: this model did not adhere to any specific paper configuration, it was tuned for reasonable training times and reduced frequency of self-attention blocks. Recipe details: * Based on [ResNet Strikes Back](https://arxiv.org/abs/2110.00476) `C` recipes * SGD (w/ Nesterov) optimizer and AGC (adaptive gradient clipping). * Cosine LR schedule with warmup This model architecture is implemented using `timm`'s flexible [BYOBNet (Bring-Your-Own-Blocks Network)](https://github.com/huggingface/pytorch-image-models/blob/main/timm/models/byobnet.py). BYOB (with BYOANet attention specific blocks) allows configuration of: * block / stage layout * block-type interleaving * stem layout * output stride (dilation) * activation and norm layers * channel and spatial / self-attention layers ...and also includes `timm` features common to many other architectures, including: * stochastic depth * gradient checkpointing * layer-wise LR decay * per-stage feature extraction ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 12.5 - GMACs: 3.3 - Activations (M): 12.0 - Image size: 256 x 256 - **Papers:** - Bottleneck Transformers for Visual Recognition: https://arxiv.org/abs/2101.11605 - ResNet strikes back: An improved training procedure in timm: https://arxiv.org/abs/2110.00476 - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('botnet26t_256.c1_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'botnet26t_256.c1_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 64, 128, 128]) # torch.Size([1, 256, 64, 64]) # torch.Size([1, 512, 32, 32]) # torch.Size([1, 1024, 16, 16]) # torch.Size([1, 2048, 8, 8]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'botnet26t_256.c1_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 2048, 8, 8) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ``` ```bibtex @article{Srinivas2021BottleneckTF, title={Bottleneck Transformers for Visual Recognition}, author={A. Srinivas and Tsung-Yi Lin and Niki Parmar and Jonathon Shlens and P. Abbeel and Ashish Vaswani}, journal={2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, year={2021}, pages={16514-16524} } ``` ```bibtex @inproceedings{wightman2021resnet, title={ResNet strikes back: An improved training procedure in timm}, author={Wightman, Ross and Touvron, Hugo and Jegou, Herve}, booktitle={NeurIPS 2021 Workshop on ImageNet: Past, Present, and Future} } ```

timm/rexnet_100.nav_in1k

timm

timm

image-classification

--- license: mit library_name: timm tags: - image-classification - timm datasets: - imagenet-1k --- # Model card for rexnet_100.nav_in1k A ReXNet image classification model. Pretrained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 4.8 - GMACs: 0.4 - Activations (M): 7.4 - Image size: 224 x 224 - **Papers:** - Rethinking Channel Dimensions for Efficient Model Design: https://arxiv.org/abs/2007.00992 - **Original:** https://github.com/clovaai/rexnet - **Dataset:** ImageNet-1k ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('rexnet_100.nav_in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Feature Map Extraction ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'rexnet_100.nav_in1k', pretrained=True, features_only=True, ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 for o in output: # print shape of each feature map in output # e.g.: # torch.Size([1, 16, 112, 112]) # torch.Size([1, 38, 56, 56]) # torch.Size([1, 61, 28, 28]) # torch.Size([1, 128, 14, 14]) # torch.Size([1, 185, 7, 7]) print(o.shape) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'rexnet_100.nav_in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 1280, 7, 7) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results)." |model |top1 |top5 |param_count|img_size|crop_pct| |-------------------------|------|------|-----------|--------|--------| |rexnetr_300.sw_in12k_ft_in1k|84.53 |97.252|34.81 |288 |1.0 | |rexnetr_200.sw_in12k_ft_in1k|83.164|96.648|16.52 |288 |1.0 | |rexnet_300.nav_in1k |82.772|96.232|34.71 |224 |0.875 | |rexnet_200.nav_in1k |81.652|95.668|16.37 |224 |0.875 | |rexnet_150.nav_in1k |80.308|95.174|9.73 |224 |0.875 | |rexnet_130.nav_in1k |79.478|94.68 |7.56 |224 |0.875 | |rexnet_100.nav_in1k |77.832|93.886|4.8 |224 |0.875 | ## Citation ```bibtex @misc{han2021rethinking, title={Rethinking Channel Dimensions for Efficient Model Design}, author={Dongyoon Han and Sangdoo Yun and Byeongho Heo and YoungJoon Yoo}, year={2021}, eprint={2007.00992}, archivePrefix={arXiv}, primaryClass={cs.CV} } ``` ```bibtex @misc{rw2019timm, author = {Ross Wightman}, title = {PyTorch Image Models}, year = {2019}, publisher = {GitHub}, journal = {GitHub repository}, doi = {10.5281/zenodo.4414861}, howpublished = {\url{https://github.com/huggingface/pytorch-image-models}} } ```

nerijs/pixel-art-xl

nerijs

diffusers

text-to-image

--- license: creativeml-openrail-m tags: - text-to-image - stable-diffusion - lora - diffusers base_model: stabilityai/stable-diffusion-xl-base-1.0 instance_prompt: pixel art widget: - text: pixel art, a cute corgi, simple, flat colors --- # Pixel Art XL ## Consider supporting further research on [Patreon](https://www.patreon.com/user?u=29466374) or [Twitter](https://twitter.com/nerijs)   Downscale 8 times to get pixel perfect images (use Nearest Neighbors) Use a fixed VAE to avoid artifacts (0.9 or fp16 fix) ### Need more performance? Use it with a LCM Lora! Use 8 steps and guidance scale of 1.5 1.2 Lora strength for the Pixel Art XL works better ```python from diffusers import DiffusionPipeline, LCMScheduler import torch model_id = "stabilityai/stable-diffusion-xl-base-1.0" lcm_lora_id = "latent-consistency/lcm-lora-sdxl" pipe = DiffusionPipeline.from_pretrained(model_id, variant="fp16") pipe.scheduler = LCMScheduler.from_config(pipe.scheduler.config) pipe.load_lora_weights(lcm_lora_id, adapter_name="lora") pipe.load_lora_weights("./pixel-art-xl.safetensors", adapter_name="pixel") pipe.set_adapters(["lora", "pixel"], adapter_weights=[1.0, 1.2]) pipe.to(device="cuda", dtype=torch.float16) prompt = "pixel, a cute corgi" negative_prompt = "3d render, realistic" num_images = 9 for i in range(num_images): img = pipe( prompt=prompt, negative_prompt=negative_prompt, num_inference_steps=8, guidance_scale=1.5, ).images[0] img.save(f"lcm_lora_{i}.png") ``` ### Tips: Don't use refiner Works great with only 1 text encoder No style prompt required No trigger keyword require Works great with isometric and non-isometric Works with 0.9 and 1.0 #### Changelog v1: Initial release

sentence-transformers/roberta-base-nli-stsb-mean-tokens

sentence-transformers

sentence-transformers

sentence-similarity

--- license: apache-2.0 library_name: sentence-transformers tags: - sentence-transformers - feature-extraction - sentence-similarity - transformers pipeline_tag: sentence-similarity --- **⚠️ This model is deprecated. Please don't use it as it produces sentence embeddings of low quality. You can find recommended sentence embedding models here: [SBERT.net - Pretrained Models](https://www.sbert.net/docs/pretrained_models.html)** # sentence-transformers/roberta-base-nli-stsb-mean-tokens This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search. ## Usage (Sentence-Transformers) Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed: ``` pip install -U sentence-transformers ``` Then you can use the model like this: ```python from sentence_transformers import SentenceTransformer sentences = ["This is an example sentence", "Each sentence is converted"] model = SentenceTransformer('sentence-transformers/roberta-base-nli-stsb-mean-tokens') embeddings = model.encode(sentences) print(embeddings) ``` ## Usage (HuggingFace Transformers) Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings. ```python from transformers import AutoTokenizer, AutoModel import torch #Mean Pooling - Take attention mask into account for correct averaging def mean_pooling(model_output, attention_mask): token_embeddings = model_output[0] #First element of model_output contains all token embeddings input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float() return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9) # Sentences we want sentence embeddings for sentences = ['This is an example sentence', 'Each sentence is converted'] # Load model from HuggingFace Hub tokenizer = AutoTokenizer.from_pretrained('sentence-transformers/roberta-base-nli-stsb-mean-tokens') model = AutoModel.from_pretrained('sentence-transformers/roberta-base-nli-stsb-mean-tokens') # Tokenize sentences encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt') # Compute token embeddings with torch.no_grad(): model_output = model(**encoded_input) # Perform pooling. In this case, max pooling. sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask']) print("Sentence embeddings:") print(sentence_embeddings) ``` ## Evaluation Results For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name=sentence-transformers/roberta-base-nli-stsb-mean-tokens) ## Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 128, 'do_lower_case': True}) with Transformer model: RobertaModel (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False}) ) ``` ## Citing & Authors This model was trained by [sentence-transformers](https://www.sbert.net/). If you find this model helpful, feel free to cite our publication [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://arxiv.org/abs/1908.10084): ```bibtex @inproceedings{reimers-2019-sentence-bert, title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", author = "Reimers, Nils and Gurevych, Iryna", booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", month = "11", year = "2019", publisher = "Association for Computational Linguistics", url = "http://arxiv.org/abs/1908.10084", } ```

timm/crossvit_9_240.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for crossvit_9_240.in1k A CrossViT image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 8.6 - GMACs: 1.8 - Activations (M): 9.5 - Image size: 240 x 240 - **Papers:** - CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification: https://arxiv.org/abs/2103.14899 - **Dataset:** ImageNet-1k - **Original:** https://github.com/IBM/CrossViT ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('crossvit_9_240.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'crossvit_9_240.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (torch.Size([1, 401, 128]), torch.Size([1, 197, 256])) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @inproceedings{ chen2021crossvit, title={{CrossViT: Cross-Attention Multi-Scale Vision Transformer for Image Classification}}, author={Chun-Fu (Richard) Chen and Quanfu Fan and Rameswar Panda}, booktitle={International Conference on Computer Vision (ICCV)}, year={2021} } ```

timm/coat_lite_mini.in1k

timm

timm

image-classification

--- tags: - image-classification - timm library_name: timm license: apache-2.0 datasets: - imagenet-1k --- # Model card for coat_lite_mini.in1k A CoaT (Co-Scale Conv-Attentional Transformer) image classification model. Trained on ImageNet-1k by paper authors. ## Model Details - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 11.0 - GMACs: 2.0 - Activations (M): 12.2 - Image size: 224 x 224 - **Papers:** - Co-Scale Conv-Attentional Image Transformers: https://arxiv.org/abs/2104.06399 - **Dataset:** ImageNet-1k - **Original:** https://github.com/mlpc-ucsd/CoaT ## Model Usage ### Image Classification ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model('coat_lite_mini.in1k', pretrained=True) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # unsqueeze single image into batch of 1 top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5) ``` ### Image Embeddings ```python from urllib.request import urlopen from PIL import Image import timm img = Image.open(urlopen( 'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png' )) model = timm.create_model( 'coat_lite_mini.in1k', pretrained=True, num_classes=0, # remove classifier nn.Linear ) model = model.eval() # get model specific transforms (normalization, resize) data_config = timm.data.resolve_model_data_config(model) transforms = timm.data.create_transform(**data_config, is_training=False) output = model(transforms(img).unsqueeze(0)) # output is (batch_size, num_features) shaped tensor # or equivalently (without needing to set num_classes=0) output = model.forward_features(transforms(img).unsqueeze(0)) # output is unpooled, a (1, 50, 512) shaped tensor output = model.forward_head(output, pre_logits=True) # output is a (1, num_features) shaped tensor ``` ## Model Comparison Explore the dataset and runtime metrics of this model in timm [model results](https://github.com/huggingface/pytorch-image-models/tree/main/results). ## Citation ```bibtex @InProceedings{Xu_2021_ICCV, author = {Xu, Weijian and Xu, Yifan and Chang, Tyler and Tu, Zhuowen}, title = {Co-Scale Conv-Attentional Image Transformers}, booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)}, month = {October}, year = {2021}, pages = {9981-9990} } ```