Update collated assets CSV.

resources/all_assets.csv

@@ -181,6 +181,7 @@ dataset,SlimPajama,Cerebras,"As of release, SlimPajama is the largest extensivel
 model,XVERSE,Xverse,XVERSE is a multilingual large language model for over 40 languages.,2023-11-06,https://github.com/xverse-ai/XVERSE-65B,https://huggingface.co/xverse/XVERSE-65B,text; text,Evaluated across a range of standard datasets regarding multiple model capabilities like language comprehension and logical reasoning.,65B parameters (dense),[],unknown,unknown,unknown,,open,custom,,,unknown,https://huggingface.co/xverse/XVERSE-65B/discussions,,,,,,,,,,,
 model,Nucleus,Nucleus.AI,Nucleus is a 22B parameters causal decoder-only model built by Nucleus.AI and trained on 500B tokens of RefinedWeb along with curated corpora.,2023-10-05,https://www.withnucleus.ai/,https://huggingface.co/NucleusAI/nucleus-22B-token-500B,text; text,"Evaluated on the OpenLLM leaderboard, performing on par with similar-sized models.",22B parameters (dense),['RefinedWeb'],unknown,2 weeks,unknown,,open,MIT,"Research on large language models; as a foundation for further specialization and finetuning for specific usecases (e.g., summarization, text generation, chatbot, etc.)",Production use without adequate assessment of risks and mitigation; any use cases which may be considered irresponsible or harmful.,unknown,https://huggingface.co/NucleusAI/nucleus-22B-token-500B/discussions,,,,,,,,,,,
 application,Cformers,Nolano,Cformers is a set of transformers that act as an API for AI inference in code.,2023-03-19,https://www.nolano.org/services/Cformers/,,,,,[],,,,,limited,MIT,,,,,,,,,,,,,,,
+model,Llama 3.1 Tulu 3,Allen Institute for AI,"Tülu3 is a leading instruction following model family, offering fully open-source data, code, and recipes designed to serve as a comprehensive guide for modern post-training techniques.",2024-11-21,https://huggingface.co/allenai/Llama-3.1-Tulu-3-8B,https://huggingface.co/allenai/Llama-3.1-Tulu-3-8B,text; text,The model can produce problematic outputs (especially when prompted to do so).,70B parameters,['Llama 3.1'],unknown,unknown,unknown,"The Tülu3 models have limited safety training, but are not deployed automatically with in-the-loop filtering of responses like ChatGPT.",limited,Llama 3.1 Community License Agreement,Tülu3 is intended for research and educational use.,The model can produce problematic outputs (especially when prompted to do so).,unknown,unknown,,,,,,,,,,,
 model,OpenAssistant LLaMA 2,OpenAssistant,OpenAssistant LLaMA 2 is an Open-Assistant fine-tuning of Meta's LLaMA 2.,2023-08-23,https://huggingface.co/OpenAssistant/llama2-70b-oasst-sft-v10,https://huggingface.co/OpenAssistant/llama2-70b-oasst-sft-v10,text; text,,70B parameters (dense),['LLaMA 2'],unknown,unknown,unknown,,open,LLaMA 2,,,unknown,https://huggingface.co/OpenAssistant/llama2-70b-oasst-sft-v10/discussions,,,,,,,,,,,
 model,ERNIE 3.0 Titan,"Baidu, PengCheng Laboratory",ERNIE 3.0 Titan is a language model,2021-12-23,https://arxiv.org/abs/2112.12731,,text; text,,260B parameters (dense),[],unknown,unknown,"Baidu V100 Cluster, PengCheng Lab Ascend 910 NPU cluster",,closed,unknown,unknown,unknown,,,,,,,,,,,,,
 model,ERNIE-ViLG,Baidu,ERNIE-ViLG is a model for text-to-image generation,2021-12-31,https://arxiv.org/abs/2112.15283,,text; image,,10B parameters (dense),[],unknown,unknown,unknown,,limited,,unknown,unknown,,,,,,,,,,,,,
@@ -276,6 +277,7 @@ model,Sonic,Cartesia,"Sonic is a low-latency voice model that generates lifelike
 model,Starling,Ollama,Starling is a large language model trained by reinforcement learning from AI feedback focused on improving chatbot helpfulness.,2023-11-02,https://starling.cs.berkeley.edu/,https://huggingface.co/berkeley-nest/Starling-LM-7B-alpha,text; text,"Mainly evaluated on MT-Bench and AlpacaEval, which are GPT-4-based comparisons.",7B parameters (dense),[],unknown,unknown,unknown,,open,CC BY NC 4.0,Academic research and free commercial usage,,,https://huggingface.co/berkeley-nest/Starling-LM-7B-alpha/discussions,,,,,,,,,,,
 model,MM1,Apple,"MM1 is a family of multimodal models, including both dense variants up to 30B and mixture-of-experts (MoE) variants up to 64B.",2024-03-16,https://arxiv.org/pdf/2403.09611.pdf,,"image, text; text",Evaluated on image captioning and visual question answering across many benchmarks.,30B parameters (dense),[],unknown,unknown,unknown,,closed,unknown,,,,,,,,,,,,,,,
 model,OpenELM,Apple,"OpenELM is a family of Open-source Efficient Language Models. It uses a layer-wise scaling strategy to efficiently allocate parameters within each layer of the transformer model, leading to enhanced accuracy.",2024-04-24,https://machinelearning.apple.com/research/openelm,https://huggingface.co/apple/OpenELM-3B-Instruct,text; text,"The models were evaluated in terms of zero-shot, LLM360, and OpenLLM leaderboard results.",3B parameters,"['RefinedWeb', 'The Pile', 'RedPajama-Data', 'Dolma', 'CoreNet library']",unknown,unknown,unknown,unknown,open,Apple,To empower and enrich the open research community by providing access to state-of-the-art language models.,"No explicit prohibited uses stated, though it is noted that users should undertake thorough safety testing.",,https://huggingface.co/apple/OpenELM-3B-Instruct/discussions,,,,,,,,,,,
+model,Depth Pro,Apple,"We present a foundation model for zero-shot metric monocular depth estimation. Our model, Depth Pro, synthesizes high-resolution depth maps with unparalleled sharpness and high-frequency details... The model is fast, producing a 2.25-megapixel depth map in 0.3 seconds on a standard GPU.",2024-10-10,https://arxiv.org/pdf/2410.02073,unknown,text; depth maps,Extensive experiments analyze specific design choices and demonstrate that Depth Pro outperforms prior work along multiple dimensions.,unknown,[],unknown,unknown,V100 GPU,"dedicated evaluation metrics for boundary accuracy in estimated depth maps, and state-of-the-art focal length estimation from a single image.",open,unknown,"Zero-shot monocular depth estimation underpins a growing variety of applications, such as advanced image editing, view synthesis, and conditional image generation.",unknown,unknown,unknown,,,,,,,,,,,
 model,BiomedGPT,Lehigh University,BiomedGPT leverages self-supervision on large and diverse datasets to accept multi-modal inputs and perform a range of downstream tasks.,2023-05-26,https://arxiv.org/pdf/2305.17100.pdf,,"image, text; text",outperforms majority of preceding state-of-the-art models over 15 unique biomedical modalities.,472M parameters (dense),"['GPT-style autoregressive decoder', 'BiomedGPT biomedical datasets']",unknown,unknown,10 NVIDIA A5000 GPUs,"No specific quality control is mentioned in model training, though details on data processing and how the model was trained are provided in the paper.",open,Apache 2.0,furthering research in developing unified and generalist models for biomedicine.,,,,,,,,,,,,,,
 model,Firefly Image 2,Adobe,"Firefly Image 2 is the next generation of generative AI for imaging, bringing significant advancements to creative control and quality, including new Text to Image capabilities now available in the popular Firefly web app where 90% of users are new to Adobe products.",2023-10-10,https://firefly.adobe.com/,,text; image,,unknown,[],unknown,unknown,unknown,,closed,unknown,creative generation of digital art and images,"AI/ML training, attempting to create abusive, illegal, or confidential content.",,,,,,,,,,,,,
 model,Firefly Vector,Adobe,"Firefly Vector is the world’s first generative AI focused on producing vector graphics, bringing Adobe's vector graphic and generative AI expertise directly into Adobe Illustrator workflows with Text to Vector Graphic.",2023-10-10,https://firefly.adobe.com/,,text; vector graphic,,unknown,[],unknown,unknown,unknown,,closed,unknown,creative generation of digital art and images,"AI/ML training, attempting to create abusive, illegal, or confidential content.",,,,,,,,,,,,,