fix: pass model_id for all causal and seq2seq lms

server/text_generation_server/models/bloom.py

@@ -90,6 +90,7 @@ class BLOOMSharded(CausalLM):
 
         torch.distributed.barrier(group=self.process_group)
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/causal_lm.py

@@ -538,6 +538,7 @@ class CausalLM(Model):
                 tokenizer.add_special_tokens({"pad_token": "[PAD]"})
 
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/galactica.py

@@ -212,6 +212,7 @@ class GalacticaSharded(CausalLM):
 
         torch.distributed.barrier(group=self.process_group)
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/gpt_neox.py

@@ -65,6 +65,7 @@ class GPTNeoxSharded(CausalLM):
 
         torch.distributed.barrier(group=self.process_group)
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/mpt.py

@@ -90,6 +90,7 @@ class MPTSharded(CausalLM):
 
         torch.distributed.barrier(group=self.process_group)
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=False,

server/text_generation_server/models/opt.py

@@ -63,6 +63,7 @@ class OPTSharded(CausalLM):
 
         torch.distributed.barrier(group=self.process_group)
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/phi.py

@@ -60,6 +60,7 @@ class Phi(CausalLM):
         model = PhiForCausalLM(config, weights)
         torch.distributed.barrier(group=self.process_group)
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/rw.py

@@ -62,6 +62,7 @@ class RW(CausalLM):
                 tokenizer.add_special_tokens({"pad_token": "[PAD]"})
 
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/santacoder.py

@@ -62,6 +62,7 @@ class SantaCoder(CausalLM):
             )
 
         super(CausalLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/seq2seq_lm.py

@@ -575,6 +575,7 @@ class Seq2SeqLM(Model):
         tokenizer.bos_token_id = model.config.decoder_start_token_id
 
         super(Seq2SeqLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,

server/text_generation_server/models/t5.py

@@ -73,6 +73,7 @@ class T5Sharded(Seq2SeqLM):
 
         torch.distributed.barrier(group=self.process_group)
         super(Seq2SeqLM, self).__init__(
+            model_id=model_id,
             model=model,
             tokenizer=tokenizer,
             requires_padding=True,