初始化项目，由ModelHub XC社区提供模型

Model: tiny-random/llama-3.3-dim64 Source: Original Platform
2026-04-11 15:09:59 +08:00
commit be6e94e81d
8 changed files with 2284 additions and 0 deletions
--- a/README.md
+++ b/README.md
@@ -0,0 +1,103 @@
+---
+library_name: transformers
+pipeline_tag: text-generation
+inference: true
+widget:
+  - text: Hello!
+    example_title: Hello world
+    group: Python
+---
+
+This tiny model is for debugging. It is randomly initialized with the config adapted from [meta-llama/Llama-3.3-70B-Instruct](https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct).
+
+### Example usage:
+
+```python
+from transformers import pipeline
+model_id = "tiny-random/llama-3.3-dim64"
+pipe = pipeline(
+    "text-generation", model=model_id, device="cuda",
+    trust_remote_code=True, max_new_tokens=3,
+)
+print(pipe("Hello World!"))
+```
+
+### Codes to create this repo:
+
+```python
+import torch
+
+from transformers import (
+    AutoConfig,
+    AutoModelForCausalLM,
+    AutoTokenizer,
+    GenerationConfig,
+    pipeline,
+    set_seed,
+)
+
+source_model_id = "meta-llama/Llama-3.3-70B-Instruct"
+save_folder = "/tmp/tiny-random/llama-3.3-dim64"
+
+tokenizer = AutoTokenizer.from_pretrained(
+    source_model_id, trust_remote_code=True,
+)
+tokenizer.save_pretrained(save_folder)
+
+config = AutoConfig.from_pretrained(
+    source_model_id, trust_remote_code=True,
+)
+config.hidden_size = 64
+config.intermediate_size = 128
+config.num_attention_heads = 2
+config.num_key_value_heads = 1
+config.head_dim = 32
+config.num_hidden_layers = 2
+config.tie_word_embeddings = True
+
+model = AutoModelForCausalLM.from_config(
+    config,
+    torch_dtype=torch.bfloat16,
+    trust_remote_code=True,
+)
+model.generation_config = GenerationConfig.from_pretrained(
+    source_model_id, trust_remote_code=True,
+)
+set_seed(42)
+with torch.no_grad():
+    for name, p in sorted(model.named_parameters()):
+        torch.nn.init.normal_(p, 0, 0.2)
+        print(name, p.shape)
+model.save_pretrained(save_folder)
+```
+
+### Printing the model:
+
+```text
+LlamaForCausalLM(
+  (model): LlamaModel(
+    (embed_tokens): Embedding(128256, 64)
+    (layers): ModuleList(
+      (0-1): 2 x LlamaDecoderLayer(
+        (self_attn): LlamaAttention(
+          (q_proj): Linear(in_features=64, out_features=64, bias=False)
+          (k_proj): Linear(in_features=64, out_features=32, bias=False)
+          (v_proj): Linear(in_features=64, out_features=32, bias=False)
+          (o_proj): Linear(in_features=64, out_features=64, bias=False)
+        )
+        (mlp): LlamaMLP(
+          (gate_proj): Linear(in_features=64, out_features=128, bias=False)
+          (up_proj): Linear(in_features=64, out_features=128, bias=False)
+          (down_proj): Linear(in_features=128, out_features=64, bias=False)
+          (act_fn): SiLU()
+        )
+        (input_layernorm): LlamaRMSNorm((64,), eps=1e-05)
+        (post_attention_layernorm): LlamaRMSNorm((64,), eps=1e-05)
+      )
+    )
+    (norm): LlamaRMSNorm((64,), eps=1e-05)
+    (rotary_emb): LlamaRotaryEmbedding()
+  )
+  (lm_head): Linear(in_features=64, out_features=128256, bias=False)
+)
+```