Push model using huggingface_hub.

README.md

@@ -25,7 +25,7 @@ You can then generate text as follows:
 ```python
 from transformers import pipeline
 
-generator = pipeline("text-generation", model="DarshanDeshpande//tmp/SLURM_20252742/tmp8za0h0al/DarshanDeshpande/gemma_2b_oasst1_ppo_model")
+generator = pipeline("text-generation", model="DarshanDeshpande//tmp/SLURM_20253966/tmph2bl7syr/DarshanDeshpande/gemma_2b_oasst1_ppo_model")
 outputs = generator("Hello, my llama is cute")
 ```
 
@@ -35,8 +35,8 @@ If you want to use the model for training or to obtain the outputs from the valu
 from transformers import AutoTokenizer
 from trl import AutoModelForCausalLMWithValueHead
 
-tokenizer = AutoTokenizer.from_pretrained("DarshanDeshpande//tmp/SLURM_20252742/tmp8za0h0al/DarshanDeshpande/gemma_2b_oasst1_ppo_model")
-model = AutoModelForCausalLMWithValueHead.from_pretrained("DarshanDeshpande//tmp/SLURM_20252742/tmp8za0h0al/DarshanDeshpande/gemma_2b_oasst1_ppo_model")
+tokenizer = AutoTokenizer.from_pretrained("DarshanDeshpande//tmp/SLURM_20253966/tmph2bl7syr/DarshanDeshpande/gemma_2b_oasst1_ppo_model")
+model = AutoModelForCausalLMWithValueHead.from_pretrained("DarshanDeshpande//tmp/SLURM_20253966/tmph2bl7syr/DarshanDeshpande/gemma_2b_oasst1_ppo_model")
 
 inputs = tokenizer("Hello, my llama is cute", return_tensors="pt")
 outputs = model(**inputs, labels=inputs["input_ids"])

adapter_config.json

@@ -19,8 +19,8 @@
   "rank_pattern": {},
   "revision": null,
   "target_modules": [
-    "q_proj",
-    "v_proj"
+    "v_proj",
+    "q_proj"
   ],
   "task_type": "CAUSAL_LM",
   "use_dora": false,

adapter_model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:2690fbb25dd45c0050b9248acd6807fa880eabf4c656cf137058b1891a5e2f17
+oid sha256:60130f2e757132e31170ed5b40cd440c9ecd5d2351a2accc897956f2951fb4f4
 size 3695848

config.json

@@ -26,9 +26,9 @@
   "cliprange": 0.2,
   "cliprange_value": 0.2,
   "vf_coef": 0.1,
-  "batch_size": 16,
+  "batch_size": 25,
   "forward_batch_size": null,
-  "mini_batch_size": 16,
+  "mini_batch_size": 25,
   "gradient_accumulation_steps": 1,
   "world_size": 1,
   "ppo_epochs": 1,
@@ -45,7 +45,7 @@
   "whiten_rewards": false,
   "is_encoder_decoder": false,
   "is_peft_model": true,
-  "backward_batch_size": 16,
-  "global_backward_batch_size": 16,
-  "global_batch_size": 16
+  "backward_batch_size": 25,
+  "global_backward_batch_size": 25,
+  "global_batch_size": 25
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:dec772d4057432c43b26ec4d65232d7b68e31d3d3bc90c6d869ebda70fdf27f7
+oid sha256:4848be04eebc4fd14531016e726563a8291df25b605763e41bc037ec0b5e0ac9
 size 9724