-
Notifications
You must be signed in to change notification settings - Fork 12
Added ministral models #116
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Draft
ParamThakkar123
wants to merge
4
commits into
main
Choose a base branch
from
add/ministral
base: main
Could not load branches
Branch not found: {{ refName }}
Loading
Could not load tags
Nothing to show
Loading
Are you sure you want to change the base?
Some commits from the old base branch may be removed from the timeline,
and old review comments may become outdated.
+278
−0
Draft
Changes from all commits
Commits
Show all changes
4 commits
Select commit
Hold shift + click to select a range
File filter
Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
There are no files selected for viewing
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,278 @@ | ||
| [ | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-14B-Instruct-2512", | ||
| "name": "Ministral-3-14B-Instruct-2512", | ||
| "description": "Ministral 3 14B is a high-performance, instruction-tuned language model with vision capabilities, delivering frontier-level results comparable to larger models. Post-trained in FP8, it is optimized for chat and instruction-based tasks. Designed for efficient edge deployment, it runs across a wide range of hardware. It fits locally within 24GB VRAM in FP8 and even less with further quantization.", | ||
| "added": "2025-12-06", | ||
| "tags": [], | ||
| "parameters": "14B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-14B-Instruct-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 30028.735153198242, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-14B-Instruct-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-8B-Instruct-2512", | ||
| "name": "Ministral-3-8B-Instruct-2512", | ||
| "description": "Ministral 3 8B is a balanced, efficient language model with vision capabilities, designed for strong performance at a compact scale. Instruction-tuned in FP8, it excels at chat and instruction-following tasks. Built for edge deployment, it runs smoothly across diverse hardware setups. It fits locally within 12GB VRAM in FP8, with even lower requirements when further quantized.", | ||
| "added": "2025-12-06", | ||
| "tags": [], | ||
| "parameters": "8B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-8B-Instruct-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 19908.187615394592, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-8B-Instruct-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-3B-Instruct-2512", | ||
| "name": "Ministral-3-3B-Instruct-2512", | ||
| "description": "Ministral 3 3B is the smallest and most lightweight model in the family, offering efficient performance with vision capabilities. Instruction-tuned in FP8, it is well suited for chat and instruction-following tasks. Optimized for edge deployment, it runs on a wide range of hardware. It fits locally within 8GB VRAM in FP8, with lower requirements through further quantization.", | ||
| "added": "2025-12-06", | ||
| "tags": [], | ||
| "parameters": "3B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-3B-Instruct-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 8943.600494384766, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-3B-Instruct-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-14B-Reasoning-2512", | ||
| "name": "Ministral-3-14B-Reasoning-2512", | ||
| "description": "Ministral 3 14B is the largest model in the family, delivering frontier-level performance comparable to much larger models. Post-trained for reasoning, it excels in math, coding, and STEM-focused tasks. Designed for efficient edge deployment, it runs across a wide range of hardware. It fits locally within 32GB VRAM in BF16, or under 24GB RAM/VRAM when quantized.", | ||
| "added": "2025-12-06", | ||
| "tags": [ | ||
| "Reasoning" | ||
| ], | ||
| "parameters": "14B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-14B-Reasoning-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 53228.86742210388, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-14B-Reasoning-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-14B-Reasoning-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-14B-Reasoning-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-8B-Reasoning-2512", | ||
| "name": "Ministral-3-8B-Reasoning-2512", | ||
| "description": "Ministral 3 8B is a balanced model offering strong performance with efficient compute and vision capabilities. Post-trained for reasoning, it is well suited for math, coding, and STEM applications. Built for edge deployment, it runs across a wide range of hardware. It fits locally within 24GB VRAM in BF16, or under 12GB RAM/VRAM when quantized.", | ||
| "added": "2025-12-06", | ||
| "tags": [ | ||
| "Reasoning" | ||
| ], | ||
| "parameters": "8B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-8B-Reasoning-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 34052.34588909149, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-8B-Reasoning-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-8B-Reasoning-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-8B-Reasoning-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-3B-Reasoning-2512", | ||
| "name": "Ministral-3-3B-Reasoning-2512", | ||
| "description": "Ministral 3 3B is the smallest and most lightweight model in the family, delivering efficient performance with vision capabilities. Post-trained for reasoning, it excels at math, coding, and STEM-focused tasks. Designed for edge deployment, it runs on a wide range of hardware. It fits locally within 16GB VRAM in BF16, or under 8GB RAM/VRAM when quantized.", | ||
| "added": "2025-12-06", | ||
| "tags": [ | ||
| "Reasoning" | ||
| ], | ||
| "parameters": "3B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-3B-Reasoning-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 14715.864577293396, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-3B-Reasoning-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-3B-Reasoning-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-3B-Reasoning-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-14B-Base-2512", | ||
| "name": "Ministral-3-14B-Base-2512", | ||
| "description": "Ministral 3 14B is the largest model in the family, delivering frontier-level performance comparable to much larger models. Post-trained for reasoning, it is optimized for math, coding, and STEM workloads. Built for edge deployment, it runs efficiently across a wide range of hardware. It fits locally within 32GB VRAM in BF16, or under 24GB RAM/VRAM when quantized.", | ||
| "added": "2025-12-06", | ||
| "tags": [], | ||
| "parameters": "14B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-14B-Base-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 53228.86742210388, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-14B-Base-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-14B-Base-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-14B-Base-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-8B-Base-2512", | ||
| "name": "Ministral-3-8B-Base-2512", | ||
| "description": "Ministral 3 8B is a balanced, efficient language model with vision capabilities, designed for flexible deployment. This base pre-trained version is ideal for custom post-training and fine-tuning workflows. For chat and instruction use cases, Ministral 3 8B Instruct 2512 is recommended. It supports edge deployment and fits locally within 24GB VRAM in BF16, or under 12GB RAM/VRAM when quantized.", | ||
| "added": "2025-12-06", | ||
| "tags": [], | ||
| "parameters": "8B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-8B-Base-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 34052.34588909149, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-8B-Base-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-8B-Base-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-8B-Base-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| }, | ||
| { | ||
| "uniqueID": "mistralai/Ministral-3-3B-Base-2512", | ||
| "name": "Ministral-3-3B-Base-2512", | ||
| "description": "Ministral 3 3B is the smallest and most lightweight model in the family, offering efficient performance with vision capabilities. This base pre-trained version is well suited for custom fine-tuning and post-training workflows. For chat and instruction use cases, Ministral 3 3B Instruct 2512 is recommended. Designed for edge deployment, it fits within 16GB VRAM in BF16, or under 8GB RAM/VRAM when quantized.", | ||
| "added": "2025-12-06", | ||
| "tags": [], | ||
| "parameters": "3B", | ||
| "context": "2048", | ||
| "architecture": "Mistral3ForConditionalGeneration", | ||
| "formats": [ | ||
| "Safetensors" | ||
| ], | ||
| "huggingface_repo": "mistralai/Ministral-3-3B-Base-2512", | ||
| "transformers_version": "5.0.0.dev0", | ||
| "gated": false, | ||
| "license": "apache-2.0", | ||
| "logo": "https://huggingface.co/mistralai/Mistral-Small-3.1-24B-Instruct-2503", | ||
| "size_of_model_in_mb": 14715.864577293396, | ||
| "author": { | ||
| "name": "mistralai", | ||
| "url": "https://huggingface.co/mistralai/Ministral-3-3B-Base-2512", | ||
| "blurb": "" | ||
| }, | ||
| "resources": { | ||
| "canonicalUrl": "https://huggingface.co/mistralai/Ministral-3-3B-Base-2512", | ||
| "downloadUrl": "https://huggingface.co/mistralai/Ministral-3-3B-Base-2512", | ||
| "paperUrl": "?" | ||
| }, | ||
| "model_group": "mistral" | ||
| } | ||
| ] | ||
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Does this work with the version of transformers we have?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I tried it with vLLM server plugin and it worked with it, The fastchat server doesn't have support for it yet.