Quantized moondream implementation and BOS token (#1980)

* moondream implementation

* add moondream example

* change config default activation

* Add assets and integrate phi mixformer with example

* Make use of kv cache and fix seq_len bug; Clean up example code

* Add README link to example

* Remove pos_embed scaling; Remove assets; Add to README; Expand VisionConfig

* Delete image

* Use apply instead of forward

* Pass bos token at the beginning of tensor.

* Quantize moondream.

* Forward with image bos token.

* Clippy.

* Use q4_0 quantization.

* Add pointers for sequence and tokens; Remove seq_len conditional

candle-transformers/src/models/moondream.rs

@@ -25,15 +25,15 @@ fn scaled_dot_product_attention(q: &Tensor, k: &Tensor, v: &Tensor) -> Result<Te
 
 #[derive(Debug, Clone, PartialEq, serde::Deserialize)]
 pub struct VisionConfig {
-    image_embedding_dim: usize,
-    model_dim: usize,
-    hidden_dim: usize,
-    hidden_features: usize,
-    embed_len: usize,
-    embed_dim: usize,
-    num_blocks: usize,
-    num_heads: usize,
-    act: candle_nn::Activation,
+    pub(crate) image_embedding_dim: usize,
+    pub(crate) model_dim: usize,
+    pub(crate) hidden_dim: usize,
+    pub(crate) hidden_features: usize,
+    pub(crate) embed_len: usize,
+    pub(crate) embed_dim: usize,
+    pub(crate) num_blocks: usize,
+    pub(crate) num_heads: usize,
+    pub(crate) act: candle_nn::Activation,
 }
 
 impl VisionConfig {