File size: 2,474 Bytes
afd288e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
05cdc2a
afd288e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
05cdc2a
afd288e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
---
license: apache-2.0
library_name: diffsynth
pipeline_tag: text-to-video
tags:
  - wan
  - wan2.1
  - echo-memory
  - text-to-video
base_model: Wan-AI/Wan2.1-T2V-1.3B
---

# Echo-Memory `context_k1` — complete Wan 2.1 1.3B DiT

DiffSynth-facing **full DiT** for the released Echo-Memory `context_k1` row.

- Paper: [arXiv:2606.09803](https://arxiv.org/abs/2606.09803)
- Code: [Echo-Team-Joy-Future-Academy-JD/Echo-Memory](https://github.com/Echo-Team-Joy-Future-Academy-JD/Echo-Memory)
- Research overlay (HF): [Echo-Team/Echo-Memory](https://huggingface.co/Echo-Team/Echo-Memory) `context_k1/epoch-0.safetensors`

## What this file is

`diffusion_pytorch_model-*-of-00006.safetensors` is a **complete Wan 2.1 T2V 1.3B DiT** (825 / 825 official keys), sharded for Hub upload. Load with `origin_file_pattern="diffusion_pytorch_model*.safetensors"`.

It is official [Wan-AI/Wan2.1-T2V-1.3B](https://huggingface.co/Wan-AI/Wan2.1-T2V-1.3B) DiT keys with the Echo-Memory `context_k1` overlay already merged. Research extras (`action_mlp`, `self_attn_with_action`, SSM / spatial slots) are **not** included, so the file loads like a normal Wan transformer (`strict=True` on `pipe.dit`).

T5 and VAE stay the official Wan files:

- `Wan-AI/Wan2.1-T2V-1.3B` `models_t5_umt5-xxl-enc-bf16.pth`
- `Wan-AI/Wan2.1-T2V-1.3B` `Wan2.1_VAE.pth`

## DiffSynth

```python
import torch
from diffsynth.pipelines.wan_video import WanVideoPipeline, ModelConfig
from diffsynth.utils.data import save_video

pipe = WanVideoPipeline.from_pretrained(
    torch_dtype=torch.bfloat16,
    device="cuda",
    model_configs=[
        ModelConfig(model_id="SOTAowner/Echo-Memory-context-k1", origin_file_pattern="diffusion_pytorch_model*.safetensors"),
        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-1.3B", origin_file_pattern="models_t5_umt5-xxl-enc-bf16.pth"),
        ModelConfig(model_id="Wan-AI/Wan2.1-T2V-1.3B", origin_file_pattern="Wan2.1_VAE.pth"),
    ],
)
video = pipe(prompt="A toy bear on a table, the camera rotates around it", seed=42)
save_video(video, "echo_memory_context_k1.mp4")
```

After this repo is mirrored on ModelScope, replace the DiT `model_id` with that ModelScope id (DiffSynth downloads from ModelScope by default).

## Notes

- 30,000-step `epoch-0` fine-tune, 640×352, 81-frame chunks.
- Camera-action / multi-chunk revisit protocol stays in the Echo-Memory repo; this file is the Wan DiT overlay only.
- Apache-2.0. Please cite the Echo-Memory paper if you use it.