Update README.md

this pr includes
* fix img embeddings
* better markdown for python codes

README.md

@@ -6,7 +6,7 @@ tags:
 - ColBERT
 ---
 <p align="center">
-  <img align="center" src="docs/images/colbertofficial.png" width="430px" />
+  <img align="center" src="https://github.com/stanford-futuredata/ColBERT/blob/main/docs/images/colbertofficial.png?raw=true" width="430px" />
 </p>
 <p align="left">
 
@@ -18,7 +18,7 @@ tags:
 
 
 <p align="center">
-  <img align="center" src="docs/images/ColBERT-Framework-MaxSim-W370px.png" />
+  <img align="center" src="https://github.com/stanford-futuredata/ColBERT/blob/main/docs/images/ColBERT-Framework-MaxSim-W370px.png?raw=true" />
 </p>
 <p align="center">
   <b>Figure 1:</b> ColBERT's late interaction, efficiently scoring the fine-grained similarity between a queries and a passage.
@@ -107,7 +107,7 @@ For fast retrieval, indexing precomputes the ColBERT representations of passages
 
 Example usage:
 
-```
+```python
 from colbert.infra import Run, RunConfig, ColBERTConfig
 from colbert import Indexer
 
@@ -127,7 +127,7 @@ if __name__=='__main__':
 
 We typically recommend that you use ColBERT for **end-to-end** retrieval, where it directly finds its top-k passages from the full collection:
 
-```
+```python
 from colbert.data import Queries
 from colbert.infra import Run, RunConfig, ColBERTConfig
 from colbert import Searcher
@@ -161,7 +161,7 @@ Training requires a JSONL triples file with a `[qid, pid+, pid-]` list per line.
 
 Example usage (training on 4 GPUs):
 
-```
+```python
 from colbert.infra import Run, RunConfig, ColBERTConfig
 from colbert import Trainer