Instructions to use unsloth/Mistral-Medium-3.5-128B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Local Apps
- Unsloth Studio new
How to use unsloth/Mistral-Medium-3.5-128B with Unsloth Studio:
Install Unsloth Studio (macOS, Linux, WSL)
curl -fsSL https://unsloth.ai/install.sh | sh # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for unsloth/Mistral-Medium-3.5-128B to start chatting
Install Unsloth Studio (Windows)
irm https://unsloth.ai/install.ps1 | iex # Run unsloth studio unsloth studio -H 0.0.0.0 -p 8888 # Then open http://localhost:8888 in your browser # Search for unsloth/Mistral-Medium-3.5-128B to start chatting
Using HuggingFace Spaces for Unsloth
# No setup required # Open https://huggingface.co/spaces/unsloth/studio in your browser # Search for unsloth/Mistral-Medium-3.5-128B to start chatting
Load model with FastModel
pip install unsloth from unsloth import FastModel model, tokenizer = FastModel.from_pretrained( model_name="unsloth/Mistral-Medium-3.5-128B", max_seq_length=2048, )
File size: 1,975 Bytes
18767fd 172ff23 18767fd 172ff23 18767fd 02fdd45 18767fd e5ecee8 172ff23 18767fd 172ff23 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 | {
"architectures": [
"Mistral3ForConditionalGeneration"
],
"torch_dtype": "bfloat16",
"image_token_index": 10,
"model_type": "mistral3",
"multimodal_projector_bias": false,
"pad_token_id": 11,
"projector_hidden_act": "gelu",
"quantization_config": {
"activation_scheme": "static",
"dequantize": false,
"modules_to_not_convert": [
"model.vision_tower",
"model.multi_modal_projector",
"lm_head"
],
"quant_method": "fp8",
"weight_block_size": null
},
"spatial_merge_size": 2,
"text_config": {
"attention_dropout": 0.0,
"bos_token_id": 1,
"eos_token_id": 2,
"head_dim": 128,
"hidden_act": "silu",
"hidden_size": 12288,
"initializer_range": 0.02,
"intermediate_size": 28672,
"max_position_embeddings": 262144,
"model_type": "ministral3",
"num_attention_heads": 96,
"num_hidden_layers": 88,
"num_key_value_heads": 8,
"pad_token_id": 11,
"rms_norm_eps": 1e-05,
"rope_parameters": {
"beta_fast": 4.0,
"beta_slow": 1.0,
"factor": 64.0,
"llama_4_scaling_beta": 0,
"mscale": 1.0,
"mscale_all_dim": 0.0,
"original_max_position_embeddings": 4096,
"rope_theta": 1000000.0,
"rope_type": "yarn",
"type": "yarn"
},
"sliding_window": null,
"tie_word_embeddings": false,
"use_cache": true,
"vocab_size": 131072
},
"tie_word_embeddings": false,
"transformers_version": "5.8.0.dev0",
"unsloth_fixed": true,
"vision_config": {
"attention_dropout": 0.0,
"head_dim": 104,
"hidden_act": "silu",
"hidden_size": 1664,
"image_size": 1540,
"initializer_range": 0.02,
"intermediate_size": 8192,
"model_type": "pixtral",
"num_attention_heads": 16,
"num_channels": 3,
"num_hidden_layers": 48,
"patch_size": 14,
"rope_parameters": {
"rope_theta": 10000.0,
"rope_type": "default"
}
},
"vision_feature_layer": -1
} |