ComfyUI generates ugly images (in Forge everything is OK): workflow attached

### Expected Behavior

Model is this one: https://civitai.com/models/613147?modelVersionId=1293903
Image : https://civitai.com/images/52145465

Just look how beautiful the image is in Forge: I reproduced it completely (the image itself contains all parameters - just drag&drop it to PNG Info tab and then "send to txt2img", then generate).

Now let us reproduce the same in ComfyUI: the workflow: [0600.json](https://github.com/user-attachments/files/19529976/0600.json)



As you may have noticed, I've used CLIP Text Encode++ https://github.com/shiimizu/ComfyUI_smZNodes (I already found posts here about making embeddings the same, with link to this component). However, it doesn't help.

If we replace CLIP Text encode with regular one, nothing will change:

Replacing the scheduler improves the things a little bit


But still it is not close to what we see in Forge:


For me it looks like Comfy is doing something very much different from what A1111 is doing, and therefore, the image is worse. And this is not the only image: I am facing with this constantly.

### Actual Behavior

Ugly images produced in ComfyUI (while in A1111 everything is all right).

### Steps to Reproduce

See above.

### Debug Logs

```powershell
epth of field, dynamic angle, photo realistic:1.4, realistic skin:1.4, fashion photography, sharp, analog film grain, hyperdetailed']]]
100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 35/35 [00:06<00:00, 5.76it/s]
Prompt executed in 6.83 seconds
got prompt
[smZNodes] | smZNodes.py:204 | tokenize_with_weights tokenizer clip_l SDTokenizer SDTokenizer.tokenize_with_weights
[smZNodes] | smZNodes.py:204 | tokenize_with_weights tokenizer clip_g SDXLClipGTokenizer SDTokenizer.tokenize_with_weights
[smZNodes] | smZNodes.py:204 | encode_token_weights cond_stage_model clip_l SDClipModel ClipTokenWeightEncoder.encode_token_weights
[smZNodes] | smZNodes.py:204 | encode_token_weights cond_stage_model clip_g SDXLClipG ClipTokenWeightEncoder.encode_token_weights
[smZNodes] | smZNodes.py:248 | tokenize_with_weights tokenizer clip_l SDTokenizer ClassicTextProcessingEngine.tokenize_with_weights
[smZNodes] | smZNodes.py:248 | tokenize_with_weights tokenizer clip_g SDXLClipGTokenizer ClassicTextProcessingEngine.tokenize_with_weig
hts
[smZNodes] | smZNodes.py:204 | tokenize_with_weights tokenizer clip_l SDTokenizer SDTokenizer.tokenize_with_weights
[smZNodes] | smZNodes.py:204 | tokenize_with_weights tokenizer clip_g SDXLClipGTokenizer SDTokenizer.tokenize_with_weights
[smZNodes] | smZNodes.py:204 | encode_token_weights cond_stage_model clip_l SDClipModel ClipTokenWeightEncoder.encode_token_weights
[smZNodes] | smZNodes.py:204 | encode_token_weights cond_stage_model clip_g SDXLClipG ClipTokenWeightEncoder.encode_token_weights
[smZNodes] | smZNodes.py:248 | tokenize_with_weights tokenizer clip_l SDTokenizer ClassicTextProcessingEngine.tokenize_with_weights
[smZNodes] | smZNodes.py:248 | tokenize_with_weights tokenizer clip_g SDXLClipGTokenizer ClassicTextProcessingEngine.tokenize_with_weig
hts
100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 35/35 [00:06<00:00, 5.73it/s]
Prompt executed in 6.90 seconds
got prompt
100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 35/35 [00:06<00:00, 5.64it/s]
Prompt executed in 6.67 seconds
got prompt
Token indices sequence length is longer than the specified maximum sequence length for this model (106 > 77). Running this sequence thr
ough the model will result in indexing errors
Token indices sequence length is longer than the specified maximum sequence length for this model (106 > 77). Running this sequence thr
ough the model will result in indexing errors
100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 35/35 [00:06<00:00, 5.50it/s]
Prompt executed in 7.00 seconds
got prompt
100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 35/35 [00:11<00:00, 2.92it/s]
Prompt executed in 12.45 seconds
got prompt
Prompt executed in 0.00 seconds
got prompt
100%|██████████████████████████████████████████████████████████████████████████████████████████████████| 35/35 [00:06<00:00, 5.75it/s]
```

### Other

_No response_