{ "step": 650, "val_loss": 1.0664211511611938, "mmlu_acc": 0.3623046875, "arc_easy_acc": 0.419921875, "gsm8k_acc": 0.03125, "humaneval_acc": 0.03125, "model_config": { "sequence_len": 2048, "vocab_size": 65536, "n_layer": 20, "n_head": 10, "n_kv_head": 10, "n_embd": 1280, "bos_token": "<|bos|>", "eos_token": "<|assistant_end|>", "pad_token": "<|assistant_end|>", "chat_template": "{% if messages[0]['role'] == 'system' %}<|bos|><|user_start|>{{ messages[0]['content'] }}\n\n{{ messages[1]['content'] }}<|user_end|>{% set messages = messages[2:] %}{% else %}<|bos|>{% endif %}{% for message in messages %}{% if loop.index0 % 2 == 0 %}<|user_start|>{{ message['content'] }}<|user_end|>{% else %}<|assistant_start|>{{ message['content'] }}<|assistant_end|>{% endif %}{% endfor %}" } }