Upload DeepseekV3ForCausalLM

config.json

@@ -9,35 +9,34 @@
     "AutoModel": "DeepseekV3Model",
     "AutoModelForCausalLM": "DeepseekV3ForCausalLM"
   },
-  "aux_loss_alpha": 0.001,
-  "bos_token_id": 163584,
-  "d_ff": 64,
-  "eos_token_id": 163585,
-  "first_k_dense_replace": 1,
-  "head_dim": 64,
+  "bos_token_id": 0,
+  "eos_token_id": 1,
+  "ep_size": 1,
+  "first_k_dense_replace": 4,
+  "head_dim": 8,
   "hidden_act": "silu",
-  "hidden_size": 64,
+  "hidden_size": 128,
   "initializer_range": 0.02,
-  "intermediate_size": 64,
-  "kv_lora_rank": 512,
-  "max_position_embeddings": 131072,
+  "intermediate_size": 256,
+  "kv_lora_rank": 16,
+  "max_position_embeddings": 2048,
   "model_type": "deepseek_v3",
   "moe_intermediate_size": 64,
   "moe_layer_freq": 1,
-  "n_group": 1,
-  "n_routed_experts": 48,
+  "n_group": 2,
+  "n_routed_experts": 4,
   "n_shared_experts": 1,
   "norm_topk_prob": true,
-  "num_attention_heads": 16,
-  "num_experts_per_tok": 8,
-  "num_hidden_layers": 4,
-  "num_key_value_heads": 4,
-  "num_nextn_predict_layers": 0,
+  "num_attention_heads": 4,
+  "num_experts_per_tok": 2,
+  "num_hidden_layers": 6,
+  "num_key_value_heads": 2,
+  "num_nextn_predict_layers": 1,
   "pretraining_tp": 1,
-  "q_lora_rank": 48,
-  "qk_head_dim": 64,
-  "qk_nope_head_dim": 32,
-  "qk_rope_head_dim": 32,
+  "q_lora_rank": 32,
+  "qk_head_dim": 32,
+  "qk_nope_head_dim": 24,
+  "qk_rope_head_dim": 8,
   "quantization_config": {
     "activation_scheme": "dynamic",
     "fmt": "e4m3",
@@ -50,25 +49,24 @@
   "rms_norm_eps": 1e-06,
   "rope_interleave": true,
   "rope_scaling": {
-    "beta_fast": 1.0,
-    "beta_slow": 1.0,
-    "factor": 32.0,
+    "beta_fast": 32,
+    "beta_slow": 1,
+    "factor": 40,
     "mscale": 1.0,
     "mscale_all_dim": 1.0,
-    "original_max_position_embeddings": 4096,
+    "original_max_position_embeddings": 2048,
     "rope_type": "yarn",
     "type": "yarn"
   },
-  "rope_theta": 50000.0,
-  "routed_scaling_factor": 2.827,
+  "rope_theta": 10000,
+  "routed_scaling_factor": 2.5,
   "scoring_func": "sigmoid",
-  "seq_aux": true,
   "tie_word_embeddings": false,
   "topk_group": 1,
   "topk_method": "noaux_tc",
   "torch_dtype": "float32",
   "transformers_version": "4.51.3",
   "use_cache": true,
-  "v_head_dim": 128,
-  "vocab_size": 163840
+  "v_head_dim": 32,
+  "vocab_size": 8000
 }

generation_config.json

@@ -1,6 +1,6 @@
 {
   "_from_model_config": true,
-  "bos_token_id": 163584,
-  "eos_token_id": 163585,
+  "bos_token_id": 0,
+  "eos_token_id": 1,
   "transformers_version": "4.51.3"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bc77d5fa0f43af6e87ad6c1696b4463d169316adce893db57ae731bc268557ca
-size 115727160
+oid sha256:0d2d38348a9b09d47ac1b1e92cdc41769a73b2673487aa60d04e03dd449114c9
+size 11520880