pd95/ apertus-mini-mlx:0.5b-bf16

132 1 week ago

Experimental MLX-backed Apertus v1.1 Mini Instruct models for custom Ollama builds. Known issue: the current Apertus v1.1 Mini artifacts can produce nonsensical chat

tools thinking 0.5b 1.5b 4b
... /
json
543b39365084 · 853B
{
"architectures": [
"ApertusForCausalLM"
],
"attention_bias": false,
"attention_dropout": 0.0,
"bos_token_id": 1,
"dtype": "bfloat16",
"eos_token_id": 68,
"hidden_act": "xielu",
"hidden_dropout": 0.0,
"hidden_size": 1024,
"initializer_range": 0.02,
"intermediate_size": 6144,
"max_position_embeddings": 4096,
"mlp_bias": false,
"model_type": "apertus",
"num_attention_heads": 16,
"num_hidden_layers": 20,
"num_key_value_heads": 4,
"pad_token_id": 10,
"post_norm": false,
"qk_norm": true,
"rms_norm_eps": 1e-05,
"rope_parameters": {
"rope_theta": 500000.0,
"rope_type": "default"
},
"rope_scaling": {
"rope_type": "linear",
"factor": 1.0
},
"rope_theta": 500000.0,
"tie_word_embeddings": true,
"transformers_version": "4.57.6",
"use_cache": false,
"vocab_size": 131072
}