初始化项目，由ModelHub XC社区提供模型

Model: francescofiamingo1/FF_3 Source: Original Platform
2026-05-31 05:07:17 +08:00
commit 2e9dc165e0
10 changed files with 100458 additions and 0 deletions
--- a/config.json
+++ b/config.json
@@ -0,0 +1,22 @@
+{
+  "architectures": ["GPT2LMHeadModel"],
+  "model_type": "gpt2",
+  "vocab_size": 50257,
+  "n_embd": 2048,
+  "n_head": 16,
+  "n_inner": 8192,
+  "n_layer": 38,
+  "n_positions": 2048,
+  "n_ctx": 2048,
+  "bos_token_id": 50256,
+  "eos_token_id": 50256,
+  "activation_function": "gelu_new",
+  "attn_pdrop": 0.0,
+  "embd_pdrop": 0.0,
+  "resid_pdrop": 0.0,
+  "layer_norm_epsilon": 1e-05,
+  "initializer_range": 0.02,
+  "torch_dtype": "bfloat16",
+  "use_cache": true,
+  "tie_word_embeddings": true
+}