Falcon embeddings (#1149) [skip docker]

* also fix multipack for falcon and add smoke tests

* make sure to handle special tokens and added tokens for lora

* fix reference to model_type

* fix tests for falcon

* fix stray typo

* fixes for smoke tests

examples/falcon/config-7b-lora.yml

@@ -60,5 +60,5 @@ fsdp:
 fsdp_config:
 special_tokens:
   pad_token: "<|endoftext|>"
-  bos_token: ">>ABSTRACT<<"
+  bos_token: "<|endoftext|>"
   eos_token: "<|endoftext|>"

examples/falcon/config-7b-qlora.yml

@@ -89,5 +89,5 @@ fsdp:
 fsdp_config:
 special_tokens:
   pad_token: "<|endoftext|>"
-  bos_token: ">>ABSTRACT<<"
+  bos_token: "<|endoftext|>"
   eos_token: "<|endoftext|>"

examples/falcon/config-7b.yml

@@ -60,5 +60,5 @@ fsdp:
 fsdp_config:
 special_tokens:
   pad_token: "<|endoftext|>"
-  bos_token: ">>ABSTRACT<<"
+  bos_token: "<|endoftext|>"
   eos_token: "<|endoftext|>"