[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"model-granite-embedding-small-english-r2":3},{"id":4,"slug":5,"name":6,"hf_id":6,"organization":7,"architecture":8,"architecture_config":9,"parameters":13,"parameters_label":14,"context_length":15,"license":16,"downloads":17,"likes":18,"status":19,"description":16,"huggingface_url":20,"created_at":21,"seo":22,"category":25,"family":28,"score":30,"variants":31,"quantizations":32,"requirements":59,"gpu_recommendations":111,"fitting_gpus":154,"community":186},506,"granite-embedding-small-english-r2","ibm-granite\u002Fgranite-embedding-small-english-r2","ibm-granite","modernbert",{"layers":10,"head_dim":11,"hidden_size":12},12,32,384,40574976,"40.6M",8192,null,4331152,75,"published","https:\u002F\u002Fhuggingface.co\u002Fibm-granite\u002Fgranite-embedding-small-english-r2","2026-08-13T18:12:57+00:00",{"title":6,"description":23,"image":16,"robots":24,"canonical_url":16},"AI 模型部署决策引擎——开源 AI 的 GPU 需求、云端价格与成本分析。","index, follow",{"name":26,"slug":27},"Embedding \u002F RAG","embedding",{"name":29,"slug":7},"Ibm-granite",0.0076,[],[33,38,43,48,52,56],{"format":34,"bits":35,"size_factor":36,"quality_loss":37},"AWQ",4,0.263,0.03,{"format":39,"bits":40,"size_factor":41,"quality_loss":42},"FP16",16,1,0,{"format":44,"bits":45,"size_factor":46,"quality_loss":47},"FP8",8,0.5,0.01,{"format":49,"bits":35,"size_factor":50,"quality_loss":51},"GGUF",0.25,0.05,{"format":53,"bits":35,"size_factor":54,"quality_loss":55},"GPTQ",0.275,0.04,{"format":57,"bits":45,"size_factor":46,"quality_loss":58},"INT8",0.02,{"FP16":60,"FP8":79,"INT8":89,"AWQ":93,"GPTQ":103,"GGUF":107},{"minimum":61,"production":69,"recommended":74},{"scenario":62,"context_length":63,"batch_size":41,"weight_gb":64,"kv_cache_gb":42,"vram_gb":65,"ram_gb":11,"disk_gb":66,"requirement_source":67,"confidence":68},"minimum",4096,0.08,1.32,39.12,"inferred",0.95,{"scenario":70,"context_length":71,"batch_size":35,"weight_gb":64,"kv_cache_gb":42,"vram_gb":72,"ram_gb":11,"disk_gb":73,"requirement_source":67,"confidence":68},"production",16384,1.8,104.12,{"scenario":75,"context_length":15,"batch_size":76,"weight_gb":64,"kv_cache_gb":42,"vram_gb":77,"ram_gb":11,"disk_gb":78,"requirement_source":67,"confidence":68},"recommended",2,1.49,65.12,{"minimum":80,"production":83,"recommended":86},{"scenario":62,"context_length":63,"batch_size":41,"weight_gb":55,"kv_cache_gb":42,"vram_gb":81,"ram_gb":11,"disk_gb":82,"requirement_source":67,"confidence":68},1.26,39.06,{"scenario":70,"context_length":71,"batch_size":35,"weight_gb":55,"kv_cache_gb":42,"vram_gb":84,"ram_gb":11,"disk_gb":85,"requirement_source":67,"confidence":68},1.74,104.06,{"scenario":75,"context_length":15,"batch_size":76,"weight_gb":55,"kv_cache_gb":42,"vram_gb":87,"ram_gb":11,"disk_gb":88,"requirement_source":67,"confidence":68},1.44,65.06,{"minimum":90,"production":91,"recommended":92},{"scenario":62,"context_length":63,"batch_size":41,"weight_gb":55,"kv_cache_gb":42,"vram_gb":81,"ram_gb":11,"disk_gb":82,"requirement_source":67,"confidence":68},{"scenario":70,"context_length":71,"batch_size":35,"weight_gb":55,"kv_cache_gb":42,"vram_gb":84,"ram_gb":11,"disk_gb":85,"requirement_source":67,"confidence":68},{"scenario":75,"context_length":15,"batch_size":76,"weight_gb":55,"kv_cache_gb":42,"vram_gb":87,"ram_gb":11,"disk_gb":88,"requirement_source":67,"confidence":68},{"minimum":94,"production":97,"recommended":100},{"scenario":62,"context_length":63,"batch_size":41,"weight_gb":58,"kv_cache_gb":42,"vram_gb":95,"ram_gb":11,"disk_gb":96,"requirement_source":67,"confidence":68},1.24,39.03,{"scenario":70,"context_length":71,"batch_size":35,"weight_gb":58,"kv_cache_gb":42,"vram_gb":98,"ram_gb":11,"disk_gb":99,"requirement_source":67,"confidence":68},1.71,104.03,{"scenario":75,"context_length":15,"batch_size":76,"weight_gb":58,"kv_cache_gb":42,"vram_gb":101,"ram_gb":11,"disk_gb":102,"requirement_source":67,"confidence":68},1.41,65.03,{"minimum":104,"production":105,"recommended":106},{"scenario":62,"context_length":63,"batch_size":41,"weight_gb":58,"kv_cache_gb":42,"vram_gb":95,"ram_gb":11,"disk_gb":96,"requirement_source":67,"confidence":68},{"scenario":70,"context_length":71,"batch_size":35,"weight_gb":58,"kv_cache_gb":42,"vram_gb":98,"ram_gb":11,"disk_gb":99,"requirement_source":67,"confidence":68},{"scenario":75,"context_length":15,"batch_size":76,"weight_gb":58,"kv_cache_gb":42,"vram_gb":101,"ram_gb":11,"disk_gb":102,"requirement_source":67,"confidence":68},{"minimum":108,"production":109,"recommended":110},{"scenario":62,"context_length":63,"batch_size":41,"weight_gb":58,"kv_cache_gb":42,"vram_gb":95,"ram_gb":11,"disk_gb":96,"requirement_source":67,"confidence":68},{"scenario":70,"context_length":71,"batch_size":35,"weight_gb":58,"kv_cache_gb":42,"vram_gb":98,"ram_gb":11,"disk_gb":99,"requirement_source":67,"confidence":68},{"scenario":75,"context_length":15,"batch_size":76,"weight_gb":58,"kv_cache_gb":42,"vram_gb":101,"ram_gb":11,"disk_gb":102,"requirement_source":67,"confidence":68},{"recommended":112,"cheapest":129,"performance":132,"min_complexity":143,"cluster_options":151,"no_match":152,"required_vram_gb":101,"candidate_count":35,"usage":153},{"gpu_model":113,"gpu_model_id":114,"gpu_count":41,"total_vram_gb":45,"hour_price":115,"monthly_cost":116,"providers":117,"bandwidth_gbps":119,"fp16_tflops":114,"interconnect_score":120,"parallel_efficiency":41,"score":121,"tier":122,"perf_metric":123,"perf_label":124,"perf_value":125,"workload":16,"cost_metric":126},"RTX 5060 Ti",24,0.0554,40.44,[118],"vast",448,40,0.5394,"cheapest","vectors\u002Fs","44800 vec\u002Fs",44800,{"effective_value":127,"effective_unit":123,"capacity_factor":128},35840,0.8,{"gpu_model":113,"gpu_model_id":114,"gpu_count":41,"total_vram_gb":45,"hour_price":115,"monthly_cost":116,"providers":130,"bandwidth_gbps":119,"fp16_tflops":114,"interconnect_score":120,"parallel_efficiency":41,"score":121,"tier":122,"perf_metric":123,"perf_label":124,"perf_value":125,"workload":16,"cost_metric":131},[118],{"effective_value":127,"effective_unit":123,"capacity_factor":128},{"gpu_model":113,"gpu_model_id":114,"gpu_count":45,"total_vram_gb":133,"hour_price":115,"monthly_cost":134,"providers":135,"bandwidth_gbps":119,"fp16_tflops":114,"interconnect_score":120,"parallel_efficiency":136,"score":137,"tier":138,"perf_metric":123,"perf_label":139,"perf_value":140,"workload":16,"cost_metric":141},64,323.54,[118],0.4,0.3635,"balanced","358400 vec\u002Fs",358400,{"effective_value":142,"effective_unit":123,"capacity_factor":128},286720,{"gpu_model":113,"gpu_model_id":114,"gpu_count":76,"total_vram_gb":40,"hour_price":115,"monthly_cost":144,"providers":145,"bandwidth_gbps":119,"fp16_tflops":114,"interconnect_score":120,"parallel_efficiency":128,"score":146,"tier":122,"perf_metric":123,"perf_label":147,"perf_value":148,"workload":16,"cost_metric":149},80.88,[118],0.4509,"89600 vec\u002Fs",89600,{"effective_value":150,"effective_unit":123,"capacity_factor":128},71680,[],false,"value",[155,162,167,175,178],{"name":156,"vram_gb":40,"cheapest_hour":157,"cost":158,"provider":118,"estimated_tps":42},"Tesla V100",0.0272,{"hourly":157,"daily":159,"monthly":160,"yearly":161},0.65,19.86,238.27,{"name":113,"vram_gb":45,"cheapest_hour":115,"cost":163,"provider":118,"estimated_tps":166},{"hourly":115,"daily":164,"monthly":116,"yearly":165},1.33,485.3,22400,{"name":168,"vram_gb":114,"cheapest_hour":169,"cost":170,"provider":118,"estimated_tps":174},"RTX 3090",0.0678,{"hourly":169,"daily":171,"monthly":172,"yearly":173},1.63,49.49,593.93,46800,{"name":176,"vram_gb":40,"cheapest_hour":169,"cost":177,"provider":118,"estimated_tps":42},"RTX 4070S Ti",{"hourly":169,"daily":171,"monthly":172,"yearly":173},{"name":179,"vram_gb":10,"cheapest_hour":180,"cost":181,"provider":118,"estimated_tps":185},"RTX 5070",0.0804,{"hourly":180,"daily":182,"monthly":183,"yearly":184},1.93,58.69,704.3,33600,{"posts_count":42,"benchmarks_count":42}]