Update README.md

README.md

@@ -46,33 +46,42 @@ Developed by [Cohere](https://cohere.com/).
 | Total parameters | 2.4B |
 | Language model | 2B parameters |
 | Vision encoder | 400M parameters; custom-trained starting from [SigLIP 2 SO400M](https://huggingface.co/google/siglip2-so400m-patch16-384) |
-| Inputs | Text and one or more images |
+| Inputs | Interleaved text and images |
 | Output | Text |
-| Languages | English, German, French, Spanish, Italian, Portuguese, Hindi, Japanese, Korean, Chinese, and Arabic |
+| Languages | English, German, French, Spanish, Italian, Portuguese, Hindi, Japanese, Korean, Chinese, Arabic, and more |
 | Tokenizer vocabulary size | 262,144 |
 | LM Backbone context window | 128K tokens |
 | Multimodal training context | 8K tokens |
 | Checkpoint precision | bfloat16 |
 | License | Apache 2.0 |
 
-The language backbone supports a 128K-token context window, but multimodal training used sequences of up to 8K tokens. Longer multimodal contexts have not yet been validated and may rely on extrapolation.
+The language backbone supports a 128K-token context window, but the validated operating range for multimodal prompts is up to 8K tokens. Longer multimodal contexts may rely on extrapolation and have not been benchmarked.
 
 ## Quickstart
 
 ### Installation
 
-North Micro Vision requires Transformers 5.15.0. Until that version is released, install Transformers from source:
+Install [PyTorch](https://pytorch.org/get-started/locally/) for your platform first. North Micro Vision requires Transformers 5.15.0, together with `accelerate` for automatic device placement and Pillow for image loading. Until Transformers 5.15.0 is released, install the runtime dependencies and Transformers from source:
 
 ```bash
+uv pip install accelerate pillow
 uv pip install "git+https://github.com/huggingface/transformers.git"
 ```
 
 Once Transformers 5.15.0 is available on PyPI, install the released package with:
 
 ```bash
-uv pip install "transformers==5.15.0"
+uv pip install accelerate pillow "transformers==5.15.0"
 ```
 
+Flash Attention 2 is optional. On supported CUDA systems, install it with:
+
+```bash
+uv pip install flash-attn --no-build-isolation
+```
+
+If you do not use `uv`, replace `uv pip` with `pip` in the commands above.
+
 ### Transformers
 
 The following example loads an image from a URL and asks the model to describe it. Prompts can interleave text with one or more images; for text-only prompts, omit the image entries.
@@ -85,22 +94,19 @@ model_id = "CohereLabs/North-Micro-Vision-Instruct"
 
 processor = AutoProcessor.from_pretrained(
     model_id,
-    trust_remote_code=True,
 )
 model = AutoModelForImageTextToText.from_pretrained(
     model_id,
     dtype="auto",
     device_map="auto",
-    trust_remote_code=True,
 )
 
 # To enable Flash Attention 2, load the model with the following settings:
-# model = AutoModelForCausalLM.from_pretrained(
+# model = AutoModelForImageTextToText.from_pretrained(
 #     model_id,
 #     dtype=torch.bfloat16,
 #     attn_implementation="flash_attention_2",
 #     device_map="auto",
-#     trust_remote_code=True,
 # )
 
 image_url = "https://cdn-uploads.huggingface.co/production/uploads/66d732effe6684fc16b12c28/Io_5OCmftsmH-n158ZtPs.png"
@@ -109,7 +115,7 @@ messages = [
         "role": "user",
         "content": [
             {"type": "image", "url": image_url},
-            {"type": "text", "text": "Describe this image."},
+            {"type": "text", "text": "What do you see?"},
         ],
     }
 ]
@@ -124,20 +130,50 @@ inputs = processor.apply_chat_template(
 
 outputs = model.generate(
     **inputs,
-    max_new_tokens=32,
+    max_new_tokens=128,
+    do_sample=True,
+    temperature=0.7,
+    top_p=0.8,
+    top_k=20,
 )
 
-input_length = inputs["input_ids"].shape[-1]
-response = processor.decode(
-    outputs[0][input_length:],
+generated_ids = [
+    output_ids[len(input_ids) :]
+    for input_ids, output_ids in zip(inputs.input_ids, outputs)
+]
+response = processor.batch_decode(
+    generated_ids,
     skip_special_tokens=True,
-)
+    clean_up_tokenization_spaces=False,
+)[0]
 print(response)
 ```
 
-For multi-image prompts, add multiple image entries to the message content before the text instruction.
+The example uses the recommended Transformers sampling settings. For deterministic output, set `do_sample=False` and omit `temperature`, `top_p`, and `top_k`.
+
+### Grounding Coordinates
+
+Bounding boxes are returned as `[x1, y1, x2, y2]` on a normalized 0–1000 scale. Map them back to the original image by scaling each axis:
+
+```python
+x1_px = x1 / 1000 * image_width
+y1_px = y1 / 1000 * image_height
+x2_px = x2 / 1000 * image_width
+y2_px = y2 / 1000 * image_height
+```
+
+### vLLM
 
-Bounding coordinates are returned in the [x1,x2,y1,y2] format on a normalized 0-1000 scale and should be mapped back to the original image dimensions.
+Public vLLM support is coming soon. Until it is available, use Transformers as shown above. The recommended vLLM settings will be:
+
+```python
+temperature = 0.7
+top_p = 0.8
+top_k = 20
+min_p = 0.0
+presence_penalty = 1.5
+repetition_penalty = 1.0
+```
 
 ## Intended Use
 
@@ -154,7 +190,7 @@ North Micro Vision Instruct is intended for research and development use cases s
 - The model is intended as a compact foundation for customization rather than a replacement for larger general-purpose chat assistants.
 - It is not a reasoning model and has limited math and code-generation capabilities.
 - Tool calling and agentic workflows are not supported.
-- The model was trained without system prompts.
+- System prompts are not recommended because the model was not trained with them, although the chat template accepts the `system` role.
 - Multimodal training used an 8K-token context; longer contexts have not been validated.
 - Native-resolution inputs can increase memory use and latency as image dimensions grow.
 
@@ -472,19 +508,6 @@ The complete comparison is provided below. We ran vision-language and text-only
 
 
 
-### vLLM
-
-Public vLLM support is coming soon. For both text-only and vision-language inference, use:
-
-```python
-temperature = 0.7
-top_p = 0.8
-top_k = 20
-min_p = 0.0
-presence_penalty = 1.5
-repetition_penalty = 1.0
-```
-
 ## Citation
 
 ```bibtex