Upload folder using huggingface_hub
Browse files- config.json +71 -0
- model.safetensors +3 -0
config.json
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"MiniMaxM2ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"attn_type_list": [
|
| 7 |
+
1,
|
| 8 |
+
1
|
| 9 |
+
],
|
| 10 |
+
"auto_map": {
|
| 11 |
+
"AutoConfig": "MiniMaxAI/MiniMax-M2--configuration_minimax_m2.MiniMaxM2Config",
|
| 12 |
+
"AutoModelForCausalLM": "MiniMaxAI/MiniMax-M2--modeling_minimax_m2.MiniMaxM2ForCausalLM"
|
| 13 |
+
},
|
| 14 |
+
"bos_token_id": null,
|
| 15 |
+
"dtype": "bfloat16",
|
| 16 |
+
"eos_token_id": null,
|
| 17 |
+
"head_dim": 32,
|
| 18 |
+
"hidden_act": "silu",
|
| 19 |
+
"hidden_size": 256,
|
| 20 |
+
"initializer_range": 0.02,
|
| 21 |
+
"intermediate_size": 512,
|
| 22 |
+
"layernorm_full_attention_beta": 1.0,
|
| 23 |
+
"layernorm_linear_attention_beta": 1.0,
|
| 24 |
+
"layernorm_mlp_beta": 1.0,
|
| 25 |
+
"max_position_embeddings": 196608,
|
| 26 |
+
"mlp_intermediate_size": 1024,
|
| 27 |
+
"model_type": "minimax_m2",
|
| 28 |
+
"mtp_transformer_layers": 1,
|
| 29 |
+
"num_attention_heads": 8,
|
| 30 |
+
"num_experts_per_tok": 2,
|
| 31 |
+
"num_hidden_layers": 2,
|
| 32 |
+
"num_key_value_heads": 4,
|
| 33 |
+
"num_local_experts": 4,
|
| 34 |
+
"num_mtp_modules": 3,
|
| 35 |
+
"output_router_logits": false,
|
| 36 |
+
"partial_rotary_factor": 0.5,
|
| 37 |
+
"qk_norm_type": "per_layer",
|
| 38 |
+
"quantization_config": {
|
| 39 |
+
"activation_scheme": "dynamic",
|
| 40 |
+
"fmt": "float8_e4m3fn",
|
| 41 |
+
"quant_method": "fp8",
|
| 42 |
+
"weight_block_size": [
|
| 43 |
+
128,
|
| 44 |
+
128
|
| 45 |
+
],
|
| 46 |
+
"modules_to_not_convert": [
|
| 47 |
+
"lm_head",
|
| 48 |
+
"e_score_correction_bias",
|
| 49 |
+
"gate"
|
| 50 |
+
]
|
| 51 |
+
},
|
| 52 |
+
"rms_norm_eps": 1e-06,
|
| 53 |
+
"rope_parameters": {
|
| 54 |
+
"rope_type": "default",
|
| 55 |
+
"rope_theta": 5000000.0
|
| 56 |
+
},
|
| 57 |
+
"rope_theta": 5000000.0,
|
| 58 |
+
"rotary_dim": 16,
|
| 59 |
+
"router_aux_loss_coef": 0.001,
|
| 60 |
+
"router_jitter_noise": 0.0,
|
| 61 |
+
"scoring_func": "sigmoid",
|
| 62 |
+
"shared_intermediate_size": 0,
|
| 63 |
+
"shared_moe_mode": "sigmoid",
|
| 64 |
+
"sliding_window": null,
|
| 65 |
+
"tie_word_embeddings": false,
|
| 66 |
+
"use_cache": true,
|
| 67 |
+
"use_mtp": true,
|
| 68 |
+
"use_qk_norm": true,
|
| 69 |
+
"use_routing_bias": true,
|
| 70 |
+
"vocab_size": 200064
|
| 71 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1c72f00e6b19f0c1289f98ea04db40bf111b41fca2b739e7b0532a705c991e9f
|
| 3 |
+
size 259660512
|