Updated readme

Bumped transformers version to 4.56.1

README.md

@@ -14,6 +14,7 @@ tags:
 license: apache-2.0
 datasets:
 - tiiuae/falcon-refinedweb
+library_name: transformers
 ---
 
 # PLDR-LLM-v51-110M-4
@@ -38,7 +39,7 @@ This model is intended to be used for research purposes. Given text as input pro
 
 ### Via Huggingface Transformers Library
 
-PLDR-LLM has custom model support for Huggingface Transformers library. PLDR-LLM custom models support is developed on Transformers v4.55.4 release available at the time.
+PLDR-LLM has custom model support for Huggingface Transformers library. PLDR-LLM with custom code is evaluated on Transformers 4.56.1 available at the time.
 
 Using `pipeline`:
 ```python
@@ -47,11 +48,13 @@ from transformers import pipeline
 pipeline = pipeline(
     task="text-generation",
     model="fromthesky/PLDR-LLM-v51-110M-4",
-    device="cuda"
+    device="cuda", # or "cpu"
+    trust_remote_code=True
 )
 
-prompt="PLDR-LLM is a large language model architecture developed by Fromthesky Research Labs."
-output=pipeline(prompt, top_p=0.6, top_k=0, temperature=1, do_sample=True, max_new_tokens=100)
+prompt=('One time they had a drumming contest, and I didn’t do very well: '
+        'They said my drumming was "too intellectual"; theirs was much more pulsing.')
+output=pipeline(prompt, top_p=0.6, top_k=0, temperature=1, do_sample=True, use_cache=True, max_new_tokens=100)
 print(output[0]["generated_text"])
 ```
 
@@ -65,21 +68,24 @@ model=AutoModelForCausalLM.from_pretrained(pretrained_model_name_or_path="fromth
                                           )
 tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path="fromthesky/PLDR-LLM-v51-110M-4",
                                         add_eos_token=False,
-                                        Legacy=False,
+                                        legacy=False,
                                         trust_remote_code=True
                                        )
-prompt="PLDR-LLM is a large language model architecture developed by Fromthesky Research Labs."
+                                       
+prompt=('One time they had a drumming contest, and I didn’t do very well: '
+        'They said my drumming was "too intellectual"; theirs was much more pulsing.')
 inputs = tokenizer([prompt], return_tensors="pt").to(device=device)
 generated_ids = model.generate(**inputs,
-                                     max_new_tokens=100, 
-                                     top_p=0.6,
-                                     top_k=0, 
-                                     temperature=1, 
-                                     do_sample=True,
-                                     use_cache=True
-                                    )
+                               max_new_tokens=100, 
+                               top_p=0.6,
+                               top_k=0, 
+                               temperature=1, 
+                               do_sample=True,
+                               use_cache=True
+                              )
 print(tokenizer.decode(generated_ids[0], skip_special_tokens=True))
 ```
+<sup>\*</sup> `prompt` string is a quote from Richard Feynman in Surely You're Joking, Mr. Feynman! Adventures of a Curious Character.
 
 #### PLDR-LLM specific configurations:
 - `custom_G_type`: `None` for learned G values during pretraining, `'identity'` for LLM with SDPA equivalent, `'random'` for G values from a random normal distribution, `'external'` for custom G values that can be assigned after model initialization. This setting is more important for training purposes, for inference it is set in the model config.json file.
@@ -94,14 +100,8 @@ the output of the residual metric learner (metric tensor, **A**), output (**A<su
 See config.json for other model configuration details.
 
 #### Notes:
-- Transformers v4.55.4 causes generation with quantized cache to fail at the time of this writing.
-To overcome this issue, install the most recent updates from transformers library:
-```python
-      git clone https://github.com/huggingface/transformers
-      cd transformers
-      pip install -e ".[dev]"
-```
-We also have a fork of transformers library with PLDR-LLM model support for future development. The PLDR-LLM model files are added to the library so custom model files are not necessary. 
+- This implementation of PLDR-LLM custom code was evaluated on Transformers 4.56.1 and pytorch 2.6.0.
+- We also have a fork of transformers library with PLDR-LLM model support for future development. The PLDR-LLM model files are added to the library so custom model files are not necessary. 
 ```python
       git clone https://github.com/burcgokden/transformers
       cd transformers

config.json

@@ -34,7 +34,7 @@
   "rope_theta": 10000.0,
   "tie_word_embeddings": false,
   "torch_dtype": "float32",
-  "transformers_version": "4.55.4",
+  "transformers_version": "4.56.1",
   "use_cache": true,
   "vocab_size": 32000
 }

requirements.txt

@@ -1,4 +1,4 @@
-transformers==4.55.4
+transformers==4.56.1
 pytorch==2.6.0
 sentencepiece==0.1.99
 python==3.11