Transformers
Safetensors
llama
speculative-decoding
eagle3
draft-model
kimi-k2.5
fp8
amd-quark
quantized
no-lm-head-quantization
text-generation-inference
quark
Instructions to use amd/Kimi-K2.5-Eagle3-FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use amd/Kimi-K2.5-Eagle3-FP8 with Transformers:
# Load model directly from transformers import AutoTokenizer, LlamaForCausalLMEagle3 tokenizer = AutoTokenizer.from_pretrained("amd/Kimi-K2.5-Eagle3-FP8") model = LlamaForCausalLMEagle3.from_pretrained("amd/Kimi-K2.5-Eagle3-FP8") - Notebooks
- Google Colab
- Kaggle
| { | |
| "auto_map": { | |
| "AutoProcessor": "kimi_k25_processor.KimiK25Processor", | |
| "AutoImageProcessor": "kimi_k25_vision_processing.KimiK25VisionProcessor" | |
| }, | |
| "media_proc_cfg": { | |
| "in_patch_limit": 16384, | |
| "patch_size": 14, | |
| "image_mean": [ | |
| 0.5, | |
| 0.5, | |
| 0.5 | |
| ], | |
| "image_std": [ | |
| 0.5, | |
| 0.5, | |
| 0.5 | |
| ], | |
| "merge_kernel_size": 2, | |
| "fixed_output_tokens": null, | |
| "patch_limit_on_one_side": 512, | |
| "in_patch_limit_each_frame": 4096, | |
| "in_patch_limit_video": null, | |
| "sample_fps": 2.0, | |
| "max_num_frames_each_video": null, | |
| "temporal_merge_kernel_size": 4, | |
| "timestamp_mode": "hh:mm:ss.fff", | |
| "config_type": "media_proc.processors.moonvit.MoonViTMediaProcessorConfig" | |
| } | |
| } |