Merge pull request #112 from lehors/update-models

Update models

Author: lehors

models/BLOOM.yml

@@ -4,41 +4,59 @@ framework:
   date: '2024-12-15'
 release:
   name: BLOOM
-  version: 176B
-  date: '2024-11-15'
-  license: {  }
+  version: 366B
+  date: '2025-05-10'
   type: language
   architecture: 'transformer decoder'
   origin: ''
   producer: BigScience
   contact: ''
+  repository: ''
   huggingface: 'https://huggingface.co/bigscience/bloom'
+  license:
+    distribution:
+      name: BigScience-BLOOM-RAIL-1.0
+      path: ''
   components:
     -
       name: 'Model architecture'
       description: "Well commented code for the model's architecture"
-      license: bigscience-bloom-rail-1.0
+    -
+      name: 'Training code'
+      description: 'Code used for training the model'
+    -
+      name: 'Inference code'
+      description: 'Code used for running the model to make predictions'
+    -
+      name: 'Evaluation code'
+      description: 'Code used for evaluating the model'
     -
       name: 'Model parameters (Final)'
       description: 'Trained model parameters, weights and biases'
-      license: 'BigScience RAIL License v1.0'
+    -
+      name: 'Model parameters (Intermediate)'
+      description: 'Trained model parameters, weights and biases'
     -
       name: Datasets
       description: 'Training, validation and testing datasets used for the model'
-      license: unlicensed
+    -
+      name: 'Model metadata'
+      description: 'Any model metadata including training configuration and optimizer states'
     -
       name: 'Model card'
       description: 'Model details including performance metrics, intended use, and limitations'
-      license: unlicensed
     -
       name: 'Data card'
       description: 'Documentation for datasets including source, characteristics, and preprocessing details'
-      license: unlicensed
     -
       name: 'Technical report'
       description: 'Technical report detailing capabilities and usage instructions for the model'
-      license: unlicensed
+      license: CC-BY-4.0
     -
       name: 'Research paper'
       description: 'Research paper detailing the development and capabilities of the model'
-      license: unlicensed
+      component_path: 'https://arxiv.org/abs/2211.05100'
+      license: CC-BY-4.0
+    -
+      name: 'Evaluation results'
+      description: 'The results from evaluating the model'

models/Granite-20B-Code-Instruct.yml

@@ -5,49 +5,46 @@ framework:
 release:
   name: Granite-20B-Code-Instruct
   version: 20B
-  date: '2024-11-15'
-  license: {  }
+  date: '2025-05-09'
   type: code
   architecture: 'transformer decoder'
   origin: Granite-20B-Code-Base
   producer: IBM
   contact: ''
   repository: 'https://github.com/ibm-granite/granite-code-models'
-  huggingface: 'https://huggingface.co/ibm-granite/granite-20b-code-instruct'
+  huggingface: 'https://huggingface.co/ibm-granite/granite-20b-code-instruct-8k'
+  license:
+    distribution:
+      name: Apache-2.0
+      path: ''
   components:
     -
       name: 'Model architecture'
       description: "Well commented code for the model's architecture"
-      license: Apache-2.0
     -
       name: 'Model parameters (Final)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: 'Model parameters (Intermediate)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: Datasets
       description: 'Training, validation and testing datasets used for the model'
-      license: unlicensed
+      license: MIT
     -
       name: 'Model metadata'
       description: 'Any model metadata including training configuration and optimizer states'
-      license: unlicensed
     -
       name: 'Model card'
       description: 'Model details including performance metrics, intended use, and limitations'
-      license: unlicensed
     -
       name: 'Data card'
       description: 'Documentation for datasets including source, characteristics, and preprocessing details'
-      license: unlicensed
-    -
-      name: 'Technical report'
-      description: 'Technical report detailing capabilities and usage instructions for the model'
-      license: unlicensed
     -
       name: 'Research paper'
       description: 'Research paper detailing the development and capabilities of the model'
-      license: unlicensed
+      component_path: 'https://arxiv.org/abs/2405.04324'
+      license: CC-BY-4.0
+    -
+      name: 'Evaluation results'
+      description: 'The results from evaluating the model'

models/Granite-34B-Code-Instruct.yml

@@ -3,51 +3,48 @@ framework:
   version: '1.0'
   date: '2024-12-15'
 release:
-  name: Granite-34B-Code-Instruct
+  name: Granite-34B-Code-Instruct-8K
   version: 34B
-  date: '2024-11-15'
-  license: {  }
+  date: '2025-05-08'
   type: code
   architecture: 'transformer decoder'
   origin: Granite-34B-Code-Base
   producer: IBM
   contact: ''
   repository: 'https://github.com/ibm-granite/granite-code-models'
-  huggingface: 'https://huggingface.co/ibm-granite/granite-34b-code-instruct'
+  huggingface: 'https://huggingface.co/ibm-granite/granite-34b-code-instruct-8k'
+  license:
+    distribution:
+      name: Apache-2.0
+      path: ''
   components:
     -
       name: 'Model architecture'
       description: "Well commented code for the model's architecture"
-      license: Apache-2.0
     -
       name: 'Model parameters (Final)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: 'Model parameters (Intermediate)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: Datasets
       description: 'Training, validation and testing datasets used for the model'
-      license: unlicensed
+      license: MIT
     -
       name: 'Model metadata'
       description: 'Any model metadata including training configuration and optimizer states'
-      license: unlicensed
     -
       name: 'Model card'
       description: 'Model details including performance metrics, intended use, and limitations'
-      license: unlicensed
     -
       name: 'Data card'
       description: 'Documentation for datasets including source, characteristics, and preprocessing details'
-      license: unlicensed
-    -
-      name: 'Technical report'
-      description: 'Technical report detailing capabilities and usage instructions for the model'
-      license: unlicensed
     -
       name: 'Research paper'
       description: 'Research paper detailing the development and capabilities of the model'
-      license: unlicensed
+      component_path: 'https://arxiv.org/abs/2405.04324'
+      license: CC-BY-4.0
+    -
+      name: 'Evaluation results'
+      description: 'The results from evaluating the model'

models/Granite-3B-Code-Instruct.yml

@@ -5,49 +5,46 @@ framework:
 release:
   name: Granite-3B-Code-Instruct
   version: 3B
-  date: '2024-11-15'
-  license: {  }
+  date: '2025-05-09'
   type: code
   architecture: 'transformer decoder'
   origin: Granite-3B-Code-Base
   producer: IBM
   contact: ''
   repository: 'https://github.com/ibm-granite/granite-code-models'
-  huggingface: 'https://huggingface.co/ibm-granite/granite-3b-code-instruct'
+  huggingface: 'https://huggingface.co/ibm-granite/granite-3b-code-instruct-2k'
+  license:
+    distribution:
+      name: Apache-2.0
+      path: ''
   components:
     -
       name: 'Model architecture'
       description: "Well commented code for the model's architecture"
-      license: Apache-2.0
     -
       name: 'Model parameters (Final)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: 'Model parameters (Intermediate)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: Datasets
       description: 'Training, validation and testing datasets used for the model'
-      license: unlicensed
+      license: MIT
     -
       name: 'Model metadata'
       description: 'Any model metadata including training configuration and optimizer states'
-      license: unlicensed
     -
       name: 'Model card'
       description: 'Model details including performance metrics, intended use, and limitations'
-      license: unlicensed
     -
       name: 'Data card'
       description: 'Documentation for datasets including source, characteristics, and preprocessing details'
-      license: unlicensed
-    -
-      name: 'Technical report'
-      description: 'Technical report detailing capabilities and usage instructions for the model'
-      license: unlicensed
     -
       name: 'Research paper'
       description: 'Research paper detailing the development and capabilities of the model'
-      license: unlicensed
+      component_path: 'https://arxiv.org/abs/2405.04324'
+      license: CC-BY-4.0
+    -
+      name: 'Evaluation results'
+      description: 'The results from evaluating the model'

models/Granite-8B-Code-Instruct.yml

@@ -5,49 +5,46 @@ framework:
 release:
   name: Granite-8B-Code-Instruct
   version: 8B
-  date: '2024-11-15'
-  license: {  }
+  date: '2025-05-08'
   type: code
   architecture: 'transformer decoder'
   origin: Granite-8B-Code-Base
   producer: IBM
   contact: ''
   repository: 'https://github.com/ibm-granite/granite-code-models'
-  huggingface: 'https://huggingface.co/ibm-granite/granite-8b-code-instruct'
+  huggingface: 'https://huggingface.co/ibm-granite/granite-8b-code-instruct-4k'
+  license:
+    distribution:
+      name: Apache-2.0
+      path: ''
   components:
     -
       name: 'Model architecture'
       description: "Well commented code for the model's architecture"
-      license: Apache-2.0
     -
       name: 'Model parameters (Final)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: 'Model parameters (Intermediate)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: Datasets
       description: 'Training, validation and testing datasets used for the model'
-      license: unlicensed
+      license: MIT
     -
       name: 'Model metadata'
       description: 'Any model metadata including training configuration and optimizer states'
-      license: unlicensed
     -
       name: 'Model card'
       description: 'Model details including performance metrics, intended use, and limitations'
-      license: unlicensed
     -
       name: 'Data card'
       description: 'Documentation for datasets including source, characteristics, and preprocessing details'
-      license: unlicensed
-    -
-      name: 'Technical report'
-      description: 'Technical report detailing capabilities and usage instructions for the model'
-      license: unlicensed
     -
       name: 'Research paper'
       description: 'Research paper detailing the development and capabilities of the model'
-      license: unlicensed
+      license: CC-BY-4.0
+      component_path: 'https://arxiv.org/abs/2405.04324'
+    -
+      name: 'Evaluation results'
+      description: 'The results from evaluating the model'

models/OLMO-7B.yml

@@ -5,53 +5,56 @@ framework:
 release:
   name: OLMO-7B
   version: 7B
-  date: '2024-11-15'
-  license: {  }
+  date: '2025-05-16'
   type: language
   architecture: transformer
   origin: oLMO
   producer: 'Allen Institute'
   contact: ''
   repository: 'https://github.com/allenai/OLMo'
   huggingface: 'https://huggingface.co/allenai/OLMo-7B'
+  license:
+    distribution:
+      name: Apache-2.0
+      path: ''
   components:
     -
       name: 'Model architecture'
       description: "Well commented code for the model's architecture"
-      license: Apache-2.0
     -
       name: 'Data preprocessing code'
       description: 'Code for data cleansing, normalization, and augmentation'
-      license: Apache-2.0
     -
       name: 'Training code'
       description: 'Code used for training the model'
-      license: Apache-2.0
     -
       name: 'Inference code'
       description: 'Code used for running the model to make predictions'
-      license: Apache-2.0
     -
       name: 'Evaluation code'
       description: 'Code used for evaluating the model'
-      license: Apache-2.0
     -
       name: 'Model parameters (Final)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: 'Model parameters (Intermediate)'
       description: 'Trained model parameters, weights and biases'
-      license: Apache-2.0
     -
       name: Datasets
       description: 'Training, validation and testing datasets used for the model'
-      license: Apache-2.0
     -
       name: 'Evaluation data'
       description: 'Data used for evaluating the model'
-      license: unlicensed
+    -
+      name: 'Model card'
+      description: 'Model details including performance metrics, intended use, and limitations'
+    -
+      name: 'Data card'
+      description: 'Documentation for datasets including source, characteristics, and preprocessing details'
     -
       name: 'Research paper'
       description: 'Research paper detailing the development and capabilities of the model'
-      license: unlicensed
+      license: CC-BY-4.0
+    -
+      name: 'Evaluation results'
+      description: 'The results from evaluating the model'