Moondream: Moondream 3 Preview

{"about":"### Pratonton Moondream 3\n\nMoondream 3 Preview tergolong dalam keluarga Moondream yang berfokus pada pemahaman imej dan tugas berkaitan objek tertentu dalam adegan. Pada generasi ketiga, pembangun beralih kepada seni bina campuran pakar: jumlah 9 bilion parameter dengan kira-kira 2 bilion digunakan bagi setiap token. Pembahagian ini menggabungkan jumlah pemberat keseluruhan dengan bahagian aktif lebih kecil semasa memproses setiap token.\n\nKad rasmi Preview menerangkan 24 lapisan, pengekod visual berasaskan SigLIP dan tokenizer SuperBPE. Kemahiran khusus model asal termasuk soal jawab imej, penjanaan kapsyen, penunjukan titik dan penandaan objek dengan kotak. Ini tugas berbeza: penjelasan adegan, penerangan seluruh imej dan pengikatan ruang objek memberikan jenis hasil berlainan.\n\n### Kerja teks dalam Ailoot API\n\nBagi model ini, Ailoot API menyediakan input teks dan jawapan teks. Imej, koordinat objek dan kaedah visual khusus tidak termasuk dalam antara muka semasa. Namun, pembangun memberikan contoh khusus Preview berupa soalan tanpa imej: model menerangkan konsep berdasarkan arahan teks biasa.\n\nMod ini boleh digunakan untuk penjelasan, jawapan berdasarkan maklumat yang diberikan dan penyusunan bahan teks. Contohnya, berikan draf penerangan, tetapkan khalayak dan minta penulisan semula yang lebih jelas dengan mengekalkan fakta. Pilihan lain ialah melampirkan arahan teks dan meminta urutan tindakan dengan penjelasan ringkas setiap langkah.\n\n### Merumuskan permintaan\n\nNyatakan tugas dan format jawapan secara langsung: penjelasan ringkas, teks padu, senarai atau jadual. Untuk meneliti bahan tertentu, sertakannya dalam mesej dan tetapkan maklumat yang mesti dikekalkan. Soalan tentang adegan visual dalam mod teks memerlukan penerangan bertulis; model mengolah penerangan yang diberikan, bukan imej itu sendiri.\n\nAiloot API menyatakan konteks 32 000 token bagi Moondream 3 Preview, meliputi mesej dan bahan perbincangan semasa. Penanda Preview menunjukkan keluaran awal generasi ketiga; sifat versi ini dan versi 3.1 perlu dipertimbangkan secara berasingan, terutama keupayaan khusus model asal.","alias_of":null,"aliases":[],"author":"Moondream","canonical_model":"moondream/moondream3-preview","ctx":null,"formats":["chat","gemini","messages","responses"],"id":"moondream/moondream3-preview","input":["text"],"kind":"text","max_out":null,"methods":[{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"anthropic","fast_multiplier":1,"flex_multiplier":1,"fmt":"messages","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"Anthropic Messages","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1/messages","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"responses","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"OpenAI Responses","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1/responses","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"openai","fast_multiplier":1,"flex_multiplier":1,"fmt":"chat","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"OpenAI Chat","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1/chat/completions","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]},{"cache_read":null,"cache_write":null,"cache_write_1h":null,"cache_write_5m":null,"conditional_meter":null,"ctx":null,"family":"gemini","fast_multiplier":1,"flex_multiplier":1,"fmt":"gemini","http":"POST","img_in":null,"img_out":null,"in":0.6,"input":["text"],"label":"Gemini generateContent","max_out":null,"max_tool_calls":null,"out":5.0,"output":["text"],"path":"/v1beta/models/moondream/moondream3-preview:generateContent","request_model_id":"moondream/moondream3-preview","route_label":"Ultra","status_reads":[],"tiers":[],"tool_prices":null,"unit":"token","units":[]}],"model_aliases":[{"canonical_model_id":"moondream/moondream3-preview","request_model_id":"moondream/moondream3-preview","route_label":"Ultra"}],"name":"Moondream: Moondream 3 Preview","price":{"in":0.6,"in_label":null,"out":5.0,"out_label":null,"sort_in":0.6,"sort_out":5.0,"varies":false},"public_id":"moondream/moondream3-preview","short":"Pratonton Moondream 3 dengan seni bina MoE padat; dalam Ailoot API menjawab soalan dan arahan teks."}