Agaram 2.5B：首个从零训练、支持工具调用的泰米尔语优先模型
Agaram 2.5B: The First Tamil-First Language Model Trained from Scratch with Tool Calling. A Playbook for Under-Represented Languages
一家小团队花 1.33 万美元从零训了个泰米尔语小模型，8 张 H100 六天搞定，还开源了权重和训练细节，低资源语言玩家可以抄作业。
Murai Labs 发布 Agaram 2.5B，一个 2.5B 参数的泰米尔语-英语双语模型，用 8 张 H100 在约 6 天内从零训练，数据约 500 亿 token，其中 58% 为泰米尔语。模型经 134,428 条对话的监督微调后支持泰米尔文、拉丁化泰米尔和英文回复，并能调用工具：在 25 个工具提示上 23 次选对工具。翻译成绩在 FLORES 全部 6 个方向超过 Sarvam-1，进泰米尔语 chrF 达 54.3 对 49.2，母语泰米尔 IndicIFEval 从 10.0% 提升到 19.3%。短板是知识类问答接近随机水平，50 个日常问题中约 46% 回复被判定可用，失败几乎都是事实错误。两个模型总成本约 1.33 万美元，权重与 GGUF 文件已开源。
Agaram 2.5B: The First Tamil-First Language Model Trained from Scratch with Tool Calling. A Playbook for Under-Represented Languages
We trained Agaram 2.5B, a 2.5-billion-parameter Tamil-English language model, from scratch on about 50 billion tokens (58% Tamil, the rest English, math, code and parallel text) on one node of eight H100 GPUs in about six days, with a new tokenizer that, unlike our first model's, keeps line breaks. Supervised fine-tuning on 134,428 conversations, whose replies an open-weight teacher model wrote, and three short targeted rounds turned it into a chat model that answers in Tamil script, romanized Tamil (Tanglish) and English and can call tools: on 25 tool prompts whose schemas appear in training it picks the right tool 23 times and gives exactly the expected arguments 13 times. The released model translates better than Sarvam-1, a 2B Indic base model, in all six FLORES and IN22 directions (FLORES chrF 54.3 into Tamil and 55.2 into English, against 49.2 and 44.2), and nearly doubles its base model on native-Tamil IndicIFEval (19.3% from 10.0%). It is weak on knowledge: reading-comprehension and knowledge benchmarks sit near chance, and on 50 held-out everyday questions written by one native speaker, one LLM judge rated 46% of its replies usable (22-24 of 50 over three passes; 95% interval for 23 of 50, 33-60%), with nearly every failure a wrong fact. Our two further SFT experiments changed answer format but not the knowledge measures. We also describe our first model, Agaram 1.5B, whose failures shaped the choices here; how the judge's scores varied between passes on identical replies; and, item by item, what the targeted rounds trained against and where frontier-model text entered the data. The project ran on compute credits and our own funds, about USD 13,300 across both models; the weights are released (licence to be announced after legal review), with GGUF files for llama.cpp, at https://huggingface.co/murai-labs/agaram-2.5b-instruct; code at https://github.com/Murai-Labs/Agaram2.5b
