初始化项目，由ModelHub XC社区提供模型

Model: pathcosmos/frankenstallm
Source: Original Platform

source/eval/outputs/hf_3b_sft_best/config.json

@@ -0,0 +1,22 @@
+{
+  "architectures": [
+    "LlamaForCausalLM"
+  ],
+  "model_type": "llama",
+  "hidden_size": 3072,
+  "intermediate_size": 8192,
+  "num_hidden_layers": 28,
+  "num_attention_heads": 24,
+  "num_key_value_heads": 8,
+  "hidden_act": "silu",
+  "max_position_embeddings": 4096,
+  "initializer_range": 0.02,
+  "rms_norm_eps": 1e-05,
+  "vocab_size": 64000,
+  "rope_theta": 500000.0,
+  "rope_scaling": null,
+  "attention_bias": false,
+  "tie_word_embeddings": true,
+  "torch_dtype": "float16",
+  "transformers_version": "4.40.0"
+}

source/eval/outputs/hf_3b_sft_best/generation_config.json

@@ -0,0 +1,9 @@
+{
+  "bos_token_id": 1,
+  "eos_token_id": 2,
+  "pad_token_id": 0,
+  "max_new_tokens": 512,
+  "temperature": 0.8,
+  "top_p": 0.9,
+  "do_sample": true
+}

source/eval/outputs/hf_3b_sft_best/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a75b13d98767cb9ea11238f3a42025e00dd0d579e8229c34bc1eb24b1dfccdd2
+size 12847882992

source/eval/outputs/hf_3b_sft_best/tokenizer_config.json

@@ -0,0 +1,9 @@
+{
+  "model_type": "llama",
+  "tokenizer_class": "PreTrainedTokenizerFast",
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "unk_token": "<unk>",
+  "pad_token": "<pad>",
+  "clean_up_tokenization_spaces": false
+}