Upload DeepseekV3ForCausalLM

config.json

@@ -1,49 +1,58 @@
 {
-  "_name_or_path": "deepseekmini_pretrain/checkpoint-611314",
+  "_name_or_path": "deepseekmini_pretrain/checkpoint-107000",
   "architectures": [
-    "DeepseekV2ForCausalLM"
+    "DeepseekV3ForCausalLM"
   ],
   "attention_bias": false,
   "attention_dropout": 0.0,
   "aux_loss_alpha": 0.001,
-  "bos_token_id": 1,
-  "eos_token_id": 2,
+  "bos_token_id": 0,
+  "eos_token_id": 1,
   "ep_size": 1,
-  "first_k_dense_replace": 1,
+  "first_k_dense_replace": 3,
   "hidden_act": "silu",
   "hidden_size": 768,
   "initializer_range": 0.02,
   "intermediate_size": 2048,
-  "kv_lora_rank": 16,
-  "max_position_embeddings": 512,
-  "model_type": "deepseek_v2",
+  "kv_lora_rank": 128,
+  "max_position_embeddings": 4096,
+  "model_type": "deepseek_v3",
   "moe_intermediate_size": 512,
   "moe_layer_freq": 1,
-  "n_group": 2,
+  "n_group": 4,
   "n_routed_experts": 8,
-  "n_shared_experts": 2,
+  "n_shared_experts": 1,
   "norm_topk_prob": true,
   "num_attention_heads": 16,
-  "num_experts_per_tok": 2,
+  "num_experts_per_tok": 3,
   "num_hidden_layers": 8,
   "num_key_value_heads": 16,
-  "pad_token_id": 2,
+  "num_nextn_predict_layers": 1,
+  "pad_token_id": 1,
   "pretraining_tp": 1,
-  "q_lora_rank": 48,
+  "q_lora_rank": 256,
   "qk_nope_head_dim": 64,
   "qk_rope_head_dim": 32,
   "rms_norm_eps": 1e-06,
-  "rope_scaling": null,
+  "rope_scaling": {
+    "beta_fast": 32,
+    "beta_slow": 1,
+    "factor": 40,
+    "mscale": 1.0,
+    "mscale_all_dim": 1.0,
+    "original_max_position_embeddings": 2048,
+    "type": "yarn"
+  },
   "rope_theta": 10000.0,
-  "routed_scaling_factor": 1.0,
-  "scoring_func": "softmax",
+  "routed_scaling_factor": 2.5,
+  "scoring_func": "sigmoid",
   "seq_aux": true,
   "tie_word_embeddings": false,
-  "topk_group": 1,
-  "topk_method": "gready",
+  "topk_group": 2,
+  "topk_method": "noaux_tc",
   "torch_dtype": "float32",
-  "transformers_version": "4.48.2",
+  "transformers_version": "4.48.3",
   "use_cache": true,
   "v_head_dim": 64,
-  "vocab_size": 8192
+  "vocab_size": 16384
 }

generation_config.json

@@ -1,7 +1,7 @@
 {
   "_from_model_config": true,
-  "bos_token_id": 1,
-  "eos_token_id": 2,
-  "pad_token_id": 2,
-  "transformers_version": "4.48.2"
+  "bos_token_id": 0,
+  "eos_token_id": 1,
+  "pad_token_id": 1,
+  "transformers_version": "4.48.3"
 }

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:0f19643421a66b64d59b6ac5a6ad8c60c2d6a955c1340864247e2bb5ed9effb5
-size 430698776
+oid sha256:4931e079fdbaa22b0c91336079c7b281fc8ae648b4ad20d4664a157aec43ee0d
+size 426198272