| Spesifikasi | Nilai |
|---|---|
| Pembuat | Meta[1] |
| Model | Model teks: 1B (1.23B) dan 3B (3.21B) |
| Panjang konteks | 128K token |
| Data latihan | Sehingga 9T token, takat pengetahuan Disember 2023 |
| Lesen | Llama 3.2 Community License (tersuai, komersial) |
Daripada kad model Llama 3.2 Meta di GitHub, yang meliputi model teks sahaja 1B dan 3B (model penglihatan 11B dan 90B tiada di dalamnya); angka carta ialah untuk model bf16 yang ditala arahan. Kad itu tidak menyatakan harga API atau output maksimum. · sumber
Estimate from the version's published specs, benchmarks and reviews, 0-5; scores are only comparable within one criteria year. Scored as launched; family pin, flagship Llama 3.2 90B Vision named (also 11B vision and 1B/3B text; 25 Sep 2024). Reasoning/coding null: no AA Intelligence Index or Terminal-Bench figure at launch; Meta's benchmarks are vision/edge tables, not comparable. Context 128k (Meta blog) -> 1. Price: Meta published none; hosted price for 90B Vision Turbo was $1.20 per 1M tokens on Together (Wayback 1 Oct 2024) -> 5. Speed null. Modalities 3 for the 90B/11B (text+image in); the 1B/3B are text-only. Reviews: Simon Willison (tried the 1B), The Register (tested the 11B vision model, critical), Tom's Guide (1B/3B on a MacBook Air); no test of the 90B.[1]
I tried running my djp codebase through that tiny 1B model just now and got a surprisingly good result - by no means comprehensive, but way better than I would ever expect from a model of that size
Tasked with analyzing a basic line chart, Llama 3.2 11B's vision capabilities start to show their limitations pretty quickly.
The 1 billion parameter model runs comfortably on an M3 MacBook Air with 8GB of RAM, while the 3 billion model also works but just barely.
| Llama 3.2 · 25 Sep 2024 | 0.60 |
|---|---|
| Meta AI drops multimodal Llama 3.2 — here's why it's such a big deal · 25 Sep 2024 | 0.40 |
| Meta gives Llama 3 vision, now if only it had a brain · 6 Okt 2024 | -0.30 |
Foto daripada pembaca, berasingan daripada media bersumber pin. Nilaikan apa yang anda makan atau lihat, dan pada pin restoran, tag hidangan terus pada foto anda.
Entri pertama sentiasa sumber pin. Keyakinan keseluruhan ialah purata berpemberat bagi sejauh mana setiap rujukan menyokong masa mula dan tamat yang digunakan di atas; sesebuah rujukan diambil kira separuh kurang bagi setiap 180 hari lebih lama daripada yang terbaharu.
Siaran Meta, diterbitkan di Connect pada 25 September 2024, berbunyi 'Today, we're releasing Llama 3.2'; kad model penglihatan menyenaraikan 'Model Release Date: Sept 25, 2024', TechCrunch[4] dan VentureBeat[5] melaporkannya pada hari itu dan OpenRouter[6] menyenaraikan 1B dan 3B sebagai dicipta pada 2024-09-25.
Meta's[1] card for the text models: 1.23B and 3.21B parameters, 128k context (8k quantized), up to 9T pretraining tokens with logits from Llama 3.1 8B and 70B as targets and distillation after pruning, December 2023 cutoff, eight languages, SFT, rejection sampling and DPO post-training, 916k H100 hours, and the instruction-tuned benchmark table against Llama 3.1 8B.
Meta's[1] vision card: 10.6B and 88.8B parameters on a separately trained cross-attention vision adapter over Llama 3.1, 6B image-text pairs, 128k context, English-only for image plus text, 2.02M H100 hours, release date Sept 25, 2024, and the MMMU, ChartQA, DocVQA and VQAv2 results.
Same-day report by Kyle Wiggers: the 11B and 90B read charts, caption images and locate objects; they are not accessible in Europe because of the 'unpredictable' regulatory environment while the 1B and 3B launch there; over 350 million downloads; the 700-million-user licence clause.[1]
Report from Connect: Zuckerberg calls Llama 3.2 'our first open-source multimodal model' and open source 'the Linux of AI', Meta[1] says it is competitive with Claude 3 Haiku and GPT-4o-mini on vision and beats Gemma and Phi 3.5-mini on instruction following, with Llama usage up 10x since 3.1.
Ada yang tertinggal atau salah? Nyatakan dengan kata-kata anda sendiri: pautan yang menyokong pin ini, tarikh mula atau tamat yang berbeza beserta sebabnya, atau fakta yang tiada atau tersilap. AI menyemaknya berbanding sumber pin ini, mencari yang lebih baik, dan menambah mana-mana halaman yang menyokong anda. Sumber pin itu sendiri tetap paling diambil kira. Gambar yang menunjukkan sesuatu yang lain, atau menunjukkannya dengan buruk, turut dilihat, dan dialihkan ke bawah atau diganti.
Tajuk asal: Llama 3.2 1B, 3B, 11B and 90B Released
| Penanda aras | Llama 3.2 1B | Llama 3.2 3B | Llama 3.1 8B | Llama 3.2 11B | Llama 3.2 90B |
|---|---|---|---|---|---|
| MMLU (5-shot) | 49.3 | 63.4 | 69.4 | t/a | t/a |
| IFEval | 59.5 | 77.4 | 80.4 | t/a | t/a |
| GSM8K (CoT) | 44.4 | 77.7 | 84.5 | t/a | t/a |
| MATH (CoT) | 30.6 | 48.0 | 51.9 | t/a | t/a |
| ARC-C | 59.4 | 78.6 | 83.4 | t/a | t/a |
| BFCL V2 | 25.7 | 67.0 | 67.1 | t/a | t/a |
| MMMU (val, CoT) | t/a | t/a | t/a | 50.7 | 60.3 |
| MathVista (testmini) | t/a | t/a | t/a | 51.5 | 57.3 |
| ChartQA (test, CoT) | t/a | t/a | t/a | 83.4 | 85.5 |
| AI2 Diagram (test) | t/a | t/a | t/a | 91.1 | 92.3 |
| DocVQA (test) | t/a | t/a | t/a | 88.4 | 90.1 |
| VQAv2 (test) | t/a | t/a | t/a | 75.2 | 78.1 |