Fix model card repo names and safetensors wording

README.md

@@ -23,9 +23,9 @@ This folder is generated from `Ant-Multi-Modal-Framework/prj/M2_Encoder` and is
 - `AutoModel.from_pretrained(..., trust_remote_code=True)`
 - Zero-shot image-text retrieval and zero-shot image classification
 
-## Required Weight File
+## Included Weight File
 
-Put the model weight file in the repo root with this exact filename:
+This repo includes the model weight file in the repo root with this exact filename:
 
 `m2_encoder_1B.safetensors`
 
@@ -40,7 +40,7 @@ The original ModelScope sample computes probabilities from the raw normalized em
 ```python
 from transformers import AutoModel, AutoProcessor
 
-repo_id = "your-name/your-m2-encoder-repo"
+repo_id = "malusama/M2-Encoder-1B"
 
 model = AutoModel.from_pretrained(repo_id, trust_remote_code=True)
 processor = AutoProcessor.from_pretrained(repo_id, trust_remote_code=True)
@@ -68,7 +68,7 @@ Those logits are useful, but they are not the same computation as the raw dot pr
 Option 1:
 
 ```bash
-python upload_to_hub.py --repo-id your-name/your-m2-encoder-repo
+python upload_to_hub.py --repo-id malusama/M2-Encoder-1B
 ```
 
 Option 2:
@@ -77,7 +77,7 @@ Option 2:
 huggingface-cli login
 git init
 git lfs install
-git remote add origin https://huggingface.co/your-name/your-m2-encoder-repo
+git remote add origin https://huggingface.co/malusama/M2-Encoder-1B
 git add .
 git commit -m "Upload M2-Encoder HF export"
 git push origin main
@@ -115,4 +115,4 @@ Example response fields:
 - This is a Hugging Face remote-code adapter, not a native `transformers` implementation.
 - The underlying model code still comes from the official M2-Encoder repo.
 - You need `trust_remote_code=True`.
-- The weights are not bundled by default when exporting unless you pass `--checkpoint`.
+- The `.safetensors` weight file is already included in this Hub repo.

configuration_m2_encoder.py

@@ -25,7 +25,7 @@ class M2EncoderConfig(PretrainedConfig):
         precision=32,
         test_only=True,
         flash_attn=False,
-        model_file="m2_encoder_1B.ckpt",
+        model_file="m2_encoder_1B.safetensors",
         architectures=None,
         auto_map=None,
         **kwargs,