hf-internal-testing
/

MiniMax-M2-Small

Text Generation

Model card Files Files and versions

AntonV HF Staff commited on 2 days ago

Commit

3eb0ac9

·

verified ·

1 Parent(s): 2fa6c9d

Upload MiniMaxM2ForCausalLM

Files changed (2) hide show

config.json +6 -6
model.safetensors +2 -2

config.json CHANGED Viewed

@@ -4,20 +4,20 @@
   ],
   "attention_dropout": 0.0,
   "bos_token_id": 200034,
-  "dtype": "float32",
   "eos_token_id": 200020,
   "head_dim": 128,
   "hidden_act": "silu",
-  "hidden_size": 256,
   "initializer_range": 0.02,
-  "intermediate_size": 128,
   "max_position_embeddings": 196608,
   "model_type": "minimax_m2",
-  "num_attention_heads": 2,
   "num_experts_per_tok": 2,
   "num_hidden_layers": 12,
-  "num_key_value_heads": 1,
-  "num_local_experts": 8,
   "output_router_logits": false,
   "rms_norm_eps": 1e-06,
   "rope_parameters": {

   ],
   "attention_dropout": 0.0,
   "bos_token_id": 200034,
+  "dtype": "float16",
   "eos_token_id": 200020,
   "head_dim": 128,
   "hidden_act": "silu",
+  "hidden_size": 1024,
   "initializer_range": 0.02,
+  "intermediate_size": 2048,
   "max_position_embeddings": 196608,
   "model_type": "minimax_m2",
+  "num_attention_heads": 12,
   "num_experts_per_tok": 2,
   "num_hidden_layers": 12,
+  "num_key_value_heads": 2,
+  "num_local_experts": 16,
   "output_router_logits": false,
   "rms_norm_eps": 1e-06,
   "rope_parameters": {

model.safetensors CHANGED Viewed

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:1f85041ccc423ed70db486dd07d66537d60beeccde5b0c5d0c14e0ff25f303ee
-size 252243968

 version https://git-lfs.github.com/spec/v1
+oid sha256:52c20617f3264ca6d743d54775b804e57b71c17a770681388777f4197ecc3d54
+size 2914306800