Upload SigmaMoEForCausalLM

config.json

@@ -0,0 +1,40 @@
+{
+  "_name_or_path": "/dccstor/broccoli/huggingface/transformers/sigma_moe/wikitext/moe/fp",
+  "activation": "gelu_new",
+  "activation_after_topk": false,
+  "architectures": [
+    "SigmaMoEForCausalLM"
+  ],
+  "attention_dropout": 0.1,
+  "d_ff": 2053,
+  "d_model": 410,
+  "embd_pdrop": 0.1,
+  "expert_dropout": 0.0,
+  "expert_size": 128,
+  "initializer_range": 0.025,
+  "layer_norm_eps": 1e-05,
+  "max_position_embeddings": 512,
+  "moe_bias": false,
+  "moe_dropout": 0.0,
+  "n_experts": 16,
+  "num_attention_heads": 10,
+  "num_hidden_layers": 16,
+  "num_sparse_hidden_layers": 16,
+  "partial_rotary_factor": 0.5,
+  "qk_layernorm": false,
+  "resid_pdrop": 0.1,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "routing_regularization": 0.001,
+  "selection_mode": "sigmoid",
+  "sinkhorn_n_iters": 3,
+  "sparse_step": 1,
+  "tie_word_embeddings": false,
+  "top_k_experts": 4,
+  "torch_dtype": "float32",
+  "transformers_version": "4.37.0.dev0",
+  "use_cache": false,
+  "v_dim": null,
+  "vocab_size": 8000,
+  "weight_std_scale": 1.0
+}

generation_config.json

@@ -0,0 +1,5 @@
+{
+  "_from_model_config": true,
+  "transformers_version": "4.37.0.dev0",
+  "use_cache": false
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8c832604ed2c3d9ae84eb689f751f21a851486372a9b80a21507bf8c2622424e
+size 177436584