@@ -309,6 +309,7 @@ export const functionCallingModelIds = [
309309export const prebuiltAppConfig : AppConfig = {
310310 useIndexedDBCache : false ,
311311 model_list : [
312+ // Llama-3.2
312313 {
313314 model : "https://huggingface.co/mlc-ai/Llama-3.2-1B-Instruct-q4f32_1-MLC" ,
314315 model_id : "Llama-3.2-1B-Instruct-q4f32_1-MLC" ,
@@ -440,7 +441,92 @@ export const prebuiltAppConfig: AppConfig = {
440441 context_window_size : 4096 ,
441442 } ,
442443 } ,
443- // Hermes-2
444+ // DeepSeek-R1-Distill-Qwen
445+ // TODO(Charlie): Qwen2-1.5B is experiencing correctness issue, hence commented for now.
446+ // {
447+ // model: "https://huggingface.co/mlc-ai/DeepSeek-R1-Distill-Qwen-1.5B-q4f16_1-MLC",
448+ // model_id: "DeepSeek-R1-Distill-Qwen-1.5B-q4f16_1-MLC",
449+ // model_lib:
450+ // modelLibURLPrefix +
451+ // modelVersion +
452+ // "/Qwen2-1.5B-Instruct-q4f16_1-ctx4k_cs1k-webgpu.wasm",
453+ // low_resource_required: true,
454+ // vram_required_MB: 1629.75,
455+ // overrides: {
456+ // context_window_size: 4096,
457+ // },
458+ // },
459+ // {
460+ // model: "https://huggingface.co/mlc-ai/DeepSeek-R1-Distill-Qwen-1.5B-q4f32_1-MLC",
461+ // model_id: "DeepSeek-R1-Distill-Qwen-1.5B-q4f32_1-MLC",
462+ // model_lib:
463+ // modelLibURLPrefix +
464+ // modelVersion +
465+ // "/Qwen2-1.5B-Instruct-q4f32_1-ctx4k_cs1k-webgpu.wasm",
466+ // low_resource_required: true,
467+ // vram_required_MB: 1888.97,
468+ // overrides: {
469+ // context_window_size: 4096,
470+ // },
471+ // },
472+ {
473+ model :
474+ "https://huggingface.co/mlc-ai/DeepSeek-R1-Distill-Qwen-7B-q4f16_1-MLC" ,
475+ model_id : "DeepSeek-R1-Distill-Qwen-7B-q4f16_1-MLC" ,
476+ model_lib :
477+ modelLibURLPrefix +
478+ modelVersion +
479+ "/Qwen2-7B-Instruct-q4f16_1-ctx4k_cs1k-webgpu.wasm" ,
480+ low_resource_required : false ,
481+ vram_required_MB : 5106.67 ,
482+ overrides : {
483+ context_window_size : 4096 ,
484+ } ,
485+ } ,
486+ {
487+ model :
488+ "https://huggingface.co/mlc-ai/DeepSeek-R1-Distill-Qwen-7B-q4f32_1-MLC" ,
489+ model_id : "DeepSeek-R1-Distill-Qwen-7B-q4f32_1-MLC" ,
490+ model_lib :
491+ modelLibURLPrefix +
492+ modelVersion +
493+ "/Qwen2-7B-Instruct-q4f32_1-ctx4k_cs1k-webgpu.wasm" ,
494+ low_resource_required : false ,
495+ vram_required_MB : 5900.09 ,
496+ overrides : {
497+ context_window_size : 4096 ,
498+ } ,
499+ } ,
500+ // DeepSeek-R1-Distill-Llama
501+ {
502+ model :
503+ "https://huggingface.co/mlc-ai/DeepSeek-R1-Distill-Llama-8B-q4f32_1-MLC" ,
504+ model_id : "DeepSeek-R1-Distill-Llama-8B-q4f32_1-MLC" ,
505+ model_lib :
506+ modelLibURLPrefix +
507+ modelVersion +
508+ "/Llama-3_1-8B-Instruct-q4f32_1-ctx4k_cs1k-webgpu.wasm" ,
509+ vram_required_MB : 6101.01 ,
510+ low_resource_required : false ,
511+ overrides : {
512+ context_window_size : 4096 ,
513+ } ,
514+ } ,
515+ {
516+ model :
517+ "https://huggingface.co/mlc-ai/DeepSeek-R1-Distill-Llama-8B-q4f16_1-MLC" ,
518+ model_id : "DeepSeek-R1-Distill-Llama-8B-q4f16_1-MLC" ,
519+ model_lib :
520+ modelLibURLPrefix +
521+ modelVersion +
522+ "/Llama-3_1-8B-Instruct-q4f16_1-ctx4k_cs1k-webgpu.wasm" ,
523+ vram_required_MB : 5001.0 ,
524+ low_resource_required : false ,
525+ overrides : {
526+ context_window_size : 4096 ,
527+ } ,
528+ } ,
529+ // Hermes-3 and Hermes-2
444530 {
445531 model :
446532 "https://huggingface.co/mlc-ai/Hermes-2-Theta-Llama-3-8B-q4f16_1-MLC" ,
@@ -497,6 +583,32 @@ export const prebuiltAppConfig: AppConfig = {
497583 context_window_size : 4096 ,
498584 } ,
499585 } ,
586+ {
587+ model : "https://huggingface.co/mlc-ai/Hermes-3-Llama-3.2-3B-q4f32_1-MLC" ,
588+ model_id : "Hermes-3-Llama-3.2-3B-q4f32_1-MLC" ,
589+ model_lib :
590+ modelLibURLPrefix +
591+ modelVersion +
592+ "/Llama-3.2-3B-Instruct-q4f32_1-ctx4k_cs1k-webgpu.wasm" ,
593+ vram_required_MB : 2951.51 ,
594+ low_resource_required : true ,
595+ overrides : {
596+ context_window_size : 4096 ,
597+ } ,
598+ } ,
599+ {
600+ model : "https://huggingface.co/mlc-ai/Hermes-3-Llama-3.2-3B-q4f16_1-MLC" ,
601+ model_id : "Hermes-3-Llama-3.2-3B-q4f16_1-MLC" ,
602+ model_lib :
603+ modelLibURLPrefix +
604+ modelVersion +
605+ "/Llama-3.2-3B-Instruct-q4f16_1-ctx4k_cs1k-webgpu.wasm" ,
606+ vram_required_MB : 2263.69 ,
607+ low_resource_required : true ,
608+ overrides : {
609+ context_window_size : 4096 ,
610+ } ,
611+ } ,
500612 {
501613 model : "https://huggingface.co/mlc-ai/Hermes-3-Llama-3.1-8B-q4f32_1-MLC" ,
502614 model_id : "Hermes-3-Llama-3.1-8B-q4f32_1-MLC" ,
0 commit comments