Merge branch 'preview'

Author: DBlankvoort

Claire.yaml

@@ -36,9 +36,9 @@ datasources_basemodel:
     notes: Based on Falcon
 
 datasources_endmodel:
-    class: partial
-    link: https://github.com/bofenghuang/vigogne
-    notes: edited Alpaca
+    class: full
+    link: https://huggingface.co/datasets/OpenLLM-France/Claire-Dialogue-French-0.1
+    notes: Dataset published on HuggingFace.
 
 weights_basemodel:
     class: open
@@ -67,9 +67,9 @@ hardware_architecture:
     notes: 
 
 preprint:
-    class: closed
-    link: 
-    notes: 
+    class: partial
+    link: https://arxiv.org/pdf/2311.16840
+    notes: Preprint released for training dataset, however not for end model.
 
 paper:
     class: closed
@@ -82,9 +82,9 @@ modelcard:
     notes: 
 
 datasheet:
-    class: closed
-    link: 
-    notes: 
+    class: partial
+    link: https://huggingface.co/datasets/OpenLLM-France/Claire-Dialogue-French-0.1
+    notes: French language dataset reasonably documented.
 
 # access:
 package:
@@ -101,5 +101,4 @@ api:
 licenses:
     class: partial
     link: 
-    notes: "cc-by-nc-sa-4.0"
-
+    notes: "cc-by-nc-sa-4.0"

CodeGeeX.yaml

@@ -68,13 +68,13 @@ hardware_architecture:
 
 preprint:
     class: open
-    link: https://arxiv.org/pdf/2406.12793
-    notes: 
+    link: https://arxiv.org/abs/2303.17568v2
+    notes: Preprint published on arXiv.
 
 paper:
     class: closed
-    link: 
-    notes: 
+    link: https://dl.acm.org/doi/abs/10.1145/3580305.3599790
+    notes: Paper published in ACM.
 
 modelcard:
     class: open

CodeGen.yaml

@@ -68,8 +68,8 @@ hardware_architecture:
 
 preprint:
     class: open
-    link: https://huggingface.co/papers/2305.02309
-    notes: 
+    link: https://arxiv.org/pdf/2305.02309
+    notes: paper published on arXiv.
 
 paper:
     class: closed

CodeLlama.yaml

@@ -73,8 +73,8 @@ preprint:
 
 paper:
     class: closed
-    link: 
-    notes: No peer-reviewed paper found.
+    link: https://conf.researchr.org/details/icse-2024/llm4code-2024-papers/2/Code-Llama-Open-Foundation-Models-for-Code
+    notes: No peer-reviewed paper found. Preprint presented as keynote in ICSE24.
 
 modelcard:
     class: partial

CogView.yaml

@@ -15,7 +15,7 @@
 
 system:
     name: CogView
-    link: https://huggingface.co/THUDM/CogView4-6B
+    link: https://huggingface.co/zai-org/CogView4-6B
     type: image
     performanceclass: full
     basemodelname: CogView4-6B
@@ -26,7 +26,7 @@ system:
 
 org:
     name: Zhipu AI
-    link: https://github.com/THUDM
+    link: https://github.com/zai-org
     notes: "Zhipu AI, one of China's AI tigers."
 
 # availability:
@@ -42,17 +42,17 @@ datasources_endmodel:
 
 weights_basemodel:
     class: open
-    link: https://huggingface.co/THUDM/CogView4-6B
+    link: https://huggingface.co/zai-org/CogView4-6B
     notes: "Weights made available on HuggingFace."
 
 weights_endmodel:
     class: open
-    link: https://huggingface.co/THUDM/CogView4-6B
+    link: https://huggingface.co/zai-org/CogView4-6B
     notes: "Weights made available on HuggingFace."
 
 trainingcode:
     class: closed
-    link: https://github.com/THUDM/CogView4
+    link: https://github.com/zai-org/CogView4
     notes: Only inference code made available on GitHub.
 
 # documentation:
@@ -78,7 +78,7 @@ paper:
 
 modelcard:
     class: partial
-    link: https://huggingface.co/THUDM/CogView4-6B
+    link: https://huggingface.co/zai-org/CogView4-6B
     notes: "Model card only contains information about inference and licensing."
 
 datasheet:
@@ -100,5 +100,5 @@ api:
 
 licenses:
     class: open
-    link: https://huggingface.co/THUDM/CogView4-6B#license
+    link: https://huggingface.co/zai-org/CogView4-6B#license
     notes: "Apache-2.0, an OSI-approved license."

Cosmos.yaml

@@ -67,9 +67,9 @@ hardware_architecture:
     notes:
 
 preprint:
-    class: closed
+    class: open
     link: https://arxiv.org/abs/2501.03575
-    notes:
+    notes: preprint published on arXiv.
 
 paper:
     class: closed

DeepFloyd.yaml

@@ -74,7 +74,7 @@ preprint:
 
 paper:
     class: closed
-    link:
+    link: https://dl.acm.org/doi/10.5555/3600270.3602913
     notes: No paper found.
 
 modelcard:

DeepHermes.yaml

@@ -15,14 +15,14 @@
 
 system:
     name: DeepHermes
-    link: https://huggingface.co/NousResearch/DeepHermes-3-Llama-3-8B-Preview
+    link: https://huggingface.co/NousResearch/Hermes-4-70B
     type: text
     performanceclass: full
-    basemodelname: Llama-3.1-8B
-    endmodelname: DeepHermes-3-Llama-3-8B-Preview
+    basemodelname: Llama-3.1-70B
+    endmodelname: Hermes-4-70B
     endmodellicense: Llama 3 Community License Agreement
-    releasedate: 2025-02
-    notes: Evaluated model only available as preview. 
+    releasedate: 2025-08
+    notes: Hybrid reasoning model based on Llama.
 
 org:
     name: Nous Research
@@ -36,69 +36,69 @@ datasources_basemodel:
     notes: based on LLama3
 
 datasources_endmodel:
-    class: closed
-    link: 
-    notes: mix of synthetic data
+    class: partial
+    link: https://huggingface.co/datasets/NousResearch/Hermes-3-Dataset
+    notes: Mix of synthetic data. Data for previous model version made available.
 
 weights_basemodel:
     class: partial
-    link:  https://huggingface.co/meta-llama/Meta-Llama-3.1-8B
+    link:  https://huggingface.co/meta-llama/Meta-Llama-3.1-70B
     notes: Inspecting the training weights requires signing Meta Llama 3.1's bespoke 'community license', not an OSI recognised open license
 
 weights_endmodel:
     class: open
-    link: https://huggingface.co/NousResearch/DeepHermes-3-Llama-3-8B-Preview
-    notes: 
+    link: https://huggingface.co/NousResearch/Hermes-4-70B
+    notes: Model made available through HuggingFace.
 
 trainingcode:
-    class: closed
-    link: 
-    notes: 
+    class: full
+    link: https://github.com/NousResearch/atropos
+    notes: Training code published on GitHub.
 
 # documentation:
 code:
-    class: closed
-    link:  
-    notes: 
+    class: open
+    link: https://huggingface.co/datasets/NousResearch/Hermes-3-Dataset
+    notes: Code comprehensively documented.
 
 hardware_architecture:
-    class: closed
-    link:  
-    notes: 
+    class: open
+    link: https://arxiv.org/pdf/2508.18255
+    notes: Hardware setup comprehensively documented in paper.
 
 preprint:
-    class: closed
-    link: 
-    notes: 
+    class: open
+    link: https://arxiv.org/pdf/2508.18255
+    notes: Preprint published on arXiv.
 
 paper:
     class: closed
     link: 
-    notes: 
+    notes: No peer-reviewed paper found.
 
 modelcard:
-    class: open
-    link: https://huggingface.co/NousResearch/DeepHermes-3-Llama-3-8B-Preview
-    notes: 
+    class: partial
+    link: https://huggingface.co/NousResearch/Hermes-4-70B
+    notes: Model card primarily contains usage information.
 
 datasheet:
     class: closed
     link: 
-    notes: 
+    notes: No datasheet found.
 
 # access:
 package:
     class: closed
     link: 
-    notes: 
+    notes: No package found.
 
 api:
     class: closed
     link: 
-    notes: 
+    notes: No API found.
     metaprompt: closed
 
 licenses:
     class: closed
-    link: 
+    link: https://huggingface.co/datasets/NousResearch/Hermes-3-Dataset
     notes: Meta custom license

FLUX.1.yaml

@@ -15,11 +15,11 @@
 
 system:
     name: FLUX.1
-    link: https://huggingface.co/black-forest-labs/FLUX.1-dev
+    link: https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev
     type: image
     performanceclass: full
     basemodelname: FLUX.1-Dev
-    endmodelname: FLUX.1-Dev
+    endmodelname: FLUX.1-Krea-dev
     endmodellicense: FLUX.1 [dev] Non-Commercial License
     releasedate: 2024-07
     notes: Rectified flow image-generation model.
@@ -47,7 +47,7 @@ weights_basemodel:
 
 weights_endmodel:
     class: partial
-    link: https://huggingface.co/black-forest-labs/FLUX.1-dev
+    link: https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev
     notes: Gated model made available through HuggingFace.
 
 trainingcode:
@@ -69,8 +69,8 @@ hardware_architecture:
 
 preprint:
     class: closed
-    link:
-    notes: No preprint found.
+    link: https://www.krea.ai/blog/flux-krea-open-source-release
+    notes: No preprint found. Blog post released.
 
 paper:
     class: closed
@@ -79,7 +79,7 @@ paper:
 
 modelcard:
     class: partial
-    link: https://huggingface.co/black-forest-labs/FLUX.1-dev
+    link: https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev
     notes: Model card primarily contains inference and license information.
 
 datasheet:
@@ -101,5 +101,5 @@ api:
 
 licenses:
     class: closed
-    link: https://huggingface.co/black-forest-labs/FLUX.1-dev#license
+    link: https://huggingface.co/black-forest-labs/FLUX.1-Krea-dev#license
     notes: FLUX.1 [dev] Non-Commercial License, not an OSI-approved license.